refactor: replace history_split with current_input_file configuration

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-08 02:15:27 +08:00 · 2026-04-27 23:36:56 +08:00
parent 1e9170e385
commit 28bb85ad63
33 changed files with 184 additions and 517 deletions
--- a/API.en.md
+++ b/API.en.md
@@ -712,7 +712,7 @@ Reads runtime settings and status, including:
 - `compat` (`wide_input_strict_output`, `strip_reference_markers`)
 - `responses` / `embeddings`
 - `auto_delete` (`mode`: `none` / `single` / `all`; legacy `sessions=true` is still treated as `all`)
- `history_split` (`enabled` always returns `true`, `trigger_after_turns`)
+- `current_input_file` (`enabled` defaults to `true`, plus `min_chars`)
 - `model_aliases`
 - `env_backed`, `needs_vercel_sync`
 - `toolcall` policy is fixed to `feature_match + high` and is no longer returned or editable via settings
@@ -727,8 +727,9 @@ Hot-updates runtime settings. Supported fields:
 - `responses.store_ttl_seconds`
 - `embeddings.provider`
 - `auto_delete.mode`
- `history_split.trigger_after_turns` (`history_split.enabled` is forced on globally; legacy client writes are stored as `true`)
+- `current_input_file.enabled` / `current_input_file.min_chars`
 - `model_aliases`
+- `history_split` is retained only for legacy config compatibility and no longer affects requests
 - `toolcall` policy is fixed and is no longer writable through settings

 ### `POST /admin/settings/password`
@@ -752,9 +753,9 @@ Imports full config with:

 The request can send config directly, or wrapped as `{"config": {...}, "mode":"merge"}`.
 Query params `?mode=merge` / `?mode=replace` are also supported.
-`replace` mode replaces the full config shape while preserving Vercel sync metadata. `merge` mode merges `keys`, `api_keys`, `accounts`, and `model_aliases`, and overwrites non-empty fields under `admin`, `runtime`, `responses`, and `embeddings`. Manage `compat`, `auto_delete`, and `history_split` via `/admin/settings` or the config file; legacy `toolcall` fields are ignored.
+`replace` mode replaces the full config shape while preserving Vercel sync metadata. `merge` mode merges `keys`, `api_keys`, `accounts`, and `model_aliases`, and overwrites non-empty fields under `admin`, `runtime`, `responses`, and `embeddings`. Manage `compat`, `auto_delete`, and `current_input_file` via `/admin/settings` or the config file; `history_split` remains only for legacy compatibility; legacy `toolcall` fields are ignored.

-> Note: `merge` mode does not update `compat`, `auto_delete`, or `history_split`.
+> Note: `merge` mode does not update `compat`, `auto_delete`, or `current_input_file`.

 ### `GET /admin/config/export`

--- a/API.md
+++ b/API.md
@@ -726,7 +726,7 @@ data: {"type":"message_stop"}
 - `compat`（`wide_input_strict_output`、`strip_reference_markers`）
 - `responses` / `embeddings`
 - `auto_delete`（`mode`：`none` / `single` / `all`；旧配置 `sessions=true` 仍按 `all` 处理）
- `history_split`（`enabled` 固定返回 `true`、`trigger_after_turns`）
+- `current_input_file`（`enabled` 默认返回 `true`、`min_chars`）
 - `model_aliases`
 - `env_backed`、`needs_vercel_sync`
 - `toolcall` 策略已固定为 `feature_match + high`，不再通过 settings 返回或修改
@@ -741,8 +741,9 @@ data: {"type":"message_stop"}
 - `responses.store_ttl_seconds`
 - `embeddings.provider`
 - `auto_delete.mode`
- `history_split.trigger_after_turns`（`history_split.enabled` 已全局强制开启；旧客户端传入时会被保存为 `true`）
+- `current_input_file.enabled` / `current_input_file.min_chars`
 - `model_aliases`
+- `history_split` 仅作为旧配置兼容字段保留，不再影响请求处理
 - `toolcall` 策略已固定，不再作为可写入字段

 ### `POST /admin/settings/password`
@@ -766,9 +767,9 @@ data: {"type":"message_stop"}

 请求可直接传配置对象，或使用 `{"config": {...}, "mode":"merge"}` 包裹格式。
 也支持在查询参数里传 `?mode=merge` / `?mode=replace`。
-`replace` 模式会按完整配置结构替换（保留 Vercel 同步元信息）；`merge` 模式会合并 `keys`、`api_keys`、`accounts`、`model_aliases`，并覆盖 `admin`、`runtime`、`responses`、`embeddings` 中的非空字段。`compat`、`auto_delete`、`history_split` 建议通过 `/admin/settings` 或配置文件管理；`toolcall` 相关字段会被忽略。
+`replace` 模式会按完整配置结构替换（保留 Vercel 同步元信息）；`merge` 模式会合并 `keys`、`api_keys`、`accounts`、`model_aliases`，并覆盖 `admin`、`runtime`、`responses`、`embeddings` 中的非空字段。`compat`、`auto_delete`、`current_input_file` 建议通过 `/admin/settings` 或配置文件管理；`history_split` 仅保留为旧配置兼容字段；`toolcall` 相关字段会被忽略。

-> 注意：`merge` 模式不会更新 `compat`、`auto_delete`、`history_split`。
+> 注意：`merge` 模式不会更新 `compat`、`auto_delete`、`current_input_file`。

 ### `GET /admin/config/export`

--- a/README.MD
+++ b/README.MD
@@ -290,9 +290,9 @@ go run ./cmd/ds2api
 - `model_aliases`：OpenAI / Claude / Gemini 共用的模型 alias 映射。
 - `runtime`：账号并发、队列与 token 刷新策略，可通过 Admin Settings 热更新。
 - `auto_delete.mode`：请求结束后的远端会话清理策略，支持 `none` / `single` / `all`。
- `history_split`：轮次拆分策略；默认关闭，开启后默认从第二轮开始将旧历史上传为 `HISTORY.txt`。
- `current_input_file`：独立拆分策略；默认开启且阈值为 `0`，触发时将完整上下文合并上传为隐藏上下文文件，并跳过 `HISTORY.txt`。
- `history_split` 与 `current_input_file` 互斥，最多启用一个；两者都关闭时请求直接透传。
+- `history_split`：旧轮次拆分字段，已废弃并忽略，仅保留兼容旧配置。
+- `current_input_file`：唯一生效的独立拆分策略；默认开启且阈值为 `0`，触发时将完整上下文合并上传为隐藏上下文文件。
+- 如果关闭 `current_input_file`，请求会直接透传，不上传拆分上下文文件。
 - `thinking_injection`：默认开启；在最新 user 消息末尾追加思考增强提示词，提高高强度推理与工具调用前的思考稳定性；`prompt` 留空时使用内置默认提示词。

 环境变量完整列表见 [部署指南](docs/DEPLOY.md)，接口鉴权规则见 [API.md](API.md#鉴权规则)。
--- a/README.en.md
+++ b/README.en.md
@@ -278,7 +278,9 @@ Common fields:
 - `model_aliases`: one shared alias map for OpenAI / Claude / Gemini model names.
 - `runtime`: account concurrency, queueing, and token refresh behavior, hot-reloadable via Admin Settings.
 - `auto_delete.mode`: remote session cleanup after each request, supporting `none` / `single` / `all`.
- `history_split`: multi-turn history split policy, now forced on globally; tune its trigger threshold to avoid inlining all long history into the prompt.
+- `history_split`: legacy multi-turn history split field, now ignored and kept only for backward-compatible config loading.
+- `current_input_file`: the only active split mode; it is enabled by default and uploads the full context as a hidden context file once the character threshold is reached.
+- If you turn off `current_input_file`, requests pass through directly without uploading any split context file.

 For the full environment variable list, see [docs/DEPLOY.en.md](docs/DEPLOY.en.md). For auth behavior, see [API.en.md](API.en.md#authentication).

--- a/2
+++ b/2
@@ -1 +1 @@
-4.1.0
+4.1.1
--- a/config.example.json
+++ b/config.example.json
@@ -50,10 +50,6 @@
  "responses": {
    "store_ttl_seconds": 900
  },
-  "history_split": {
-    "enabled": false,
-    "trigger_after_turns": 1
-  },
  "current_input_file": {
    "enabled": true,
    "min_chars": 0
--- a/docs/ARCHITECTURE.en.md
+++ b/docs/ARCHITECTURE.en.md
@@ -43,7 +43,7 @@ ds2api/
 │   │       ├── responses/                # Responses API and response store
 │   │       ├── files/                    # Files API and inline-file preprocessing
 │   │       ├── embeddings/               # Embeddings API
-│   │       ├── history/                  # OpenAI history split
+│   │       ├── history/                  # OpenAI context file handling
 │   │       └── shared/                   # OpenAI HTTP errors/models/tool formatting
 │   ├── js/                               # Node runtime related logic
 │   │   ├── chat-stream/                  # Node streaming bridge
--- a/docs/ARCHITECTURE.md
+++ b/docs/ARCHITECTURE.md
@@ -43,7 +43,7 @@ ds2api/
 │   │       ├── responses/                # Responses API 与 response store
 │   │       ├── files/                    # Files API 与 inline file 预处理
 │   │       ├── embeddings/               # Embeddings API
-│   │       ├── history/                  # OpenAI history split
+│   │       ├── history/                  # OpenAI context file handling
 │   │       └── shared/                   # OpenAI HTTP 公共错误/模型/工具格式
 │   ├── js/                               # Node Runtime 相关逻辑
 │   │   ├── chat-stream/                  # Node 流式输出桥接
--- a/docs/prompt-compatibility.md
+++ b/docs/prompt-compatibility.md
@@ -3,7 +3,7 @@
 文档导航：[总览](../README.MD) / [架构说明](./ARCHITECTURE.md) / [接口文档](../API.md) / [测试指南](./TESTING.md)

 > 本文档是 DS2API“把 OpenAI / Claude / Gemini 风格 API 请求兼容成 DeepSeek 网页对话纯文本上下文”的专项说明。
-> 这是项目最重要的兼容产物之一。凡是修改消息标准化、tool prompt 注入、tool history 保留、文件引用、history split、下游 completion payload 组装等行为，都必须同步更新本文档。
+> 这是项目最重要的兼容产物之一。凡是修改消息标准化、tool prompt 注入、tool history 保留、文件引用、current input file / legacy history_split、下游 completion payload 组装等行为，都必须同步更新本文档。

 ## 1. 核心结论

@@ -45,7 +45,7 @@ DS2API 当前的核心思路，不是把客户端传来的 `messages`、`tools`
  -> promptcompat 统一消息标准化
  -> tool prompt 注入
  -> DeepSeek 风格 prompt 拼装
-  -> 文件收集 / inline 上传 / history split（OpenAI 链路）
+  -> 文件收集 / inline 上传 / current input file（OpenAI 链路）
  -> completion payload
  -> 下游网页对话接口
 ```
@@ -107,12 +107,11 @@ DS2API 当前的核心思路，不是把客户端传来的 `messages`、`tools`

 ## 5. prompt 是怎么拼出来的

-OpenAI Chat / Responses 在标准化后、history split / current input file 之前，会默认执行 `thinking_injection` 增强。它参考 DeepSeek V4 “把控制指令放在 user 消息末尾更稳定”的用法，在最新 user message 后追加思考增强提示词。当前内置默认提示词以 `Reasoning Effort: Absolute maximum with no shortcuts permitted.` 开头，并继续要求模型充分分解问题、覆盖潜在路径与边界条件、把完整推演过程显式写出。该开关默认启用，可通过 `thinking_injection.enabled=false` 关闭；也可以通过 `thinking_injection.prompt` 自定义提示词，留空时使用内置默认提示词。
+OpenAI Chat / Responses 在标准化后、current input file 之前，会默认执行 `thinking_injection` 增强。它参考 DeepSeek V4 “把控制指令放在 user 消息末尾更稳定”的用法，在最新 user message 后追加思考增强提示词。当前内置默认提示词以 `Reasoning Effort: Absolute maximum with no shortcuts permitted.` 开头，并继续要求模型充分分解问题、覆盖潜在路径与边界条件、把完整推演过程显式写出。该开关默认启用，可通过 `thinking_injection.enabled=false` 关闭；也可以通过 `thinking_injection.prompt` 自定义提示词，留空时使用内置默认提示词。

 这段增强属于 prompt 可见上下文：

 - 普通请求会直接出现在最终 `prompt` 的最新 user block 末尾。
- 如果触发 `HISTORY.txt`，它会保留在 live context 的最新 user turn 中。
 - 如果触发 current input file，它会进入完整上下文文件中。

 ### 5.1 角色标记
@@ -241,51 +240,22 @@ OpenAI 文件相关实现：

 ## 9. 多轮历史为什么不会一直完整内联在 prompt

-兼容层提供两种拆分策略：
+兼容层现在只保留 `current_input_file` 这一种拆分方式；旧的 `history_split` 已废弃，只保留为兼容旧配置的字段，不再参与请求处理。

- `history_split` 是轮次拆分，默认关闭；开启后默认从第 2 个 user turn 起触发，可通过 `history_split.trigger_after_turns` 调整阈值。
- `current_input_file` 是独立拆分，默认开启；它用于把“完整上下文”合并进隐藏上下文文件。当最新 user turn 的纯文本长度达到 `current_input_file.min_chars`（默认 `0`）时，兼容层会上传一个文件名为 `IGNORE.txt` 的上下文文件，并在 live prompt 中只保留一个中性的 user 消息要求模型直接回答最新请求，不再暴露文件名或要求模型读取本地文件。
-
-两个策略互斥，最多只能启用一个。如果两个开关都关闭，请求会直接透传，不上传 `HISTORY.txt` 或 current input file。
+- `current_input_file` 默认开启；它用于把“完整上下文”合并进隐藏上下文文件。当最新 user turn 的纯文本长度达到 `current_input_file.min_chars`（默认 `0`）时，兼容层会上传一个文件名为 `IGNORE.txt` 的上下文文件，并在 live prompt 中只保留一个中性的 user 消息要求模型直接回答最新请求，不再暴露文件名或要求模型读取本地文件。
+- 如果 `current_input_file.enabled=false`，请求会直接透传，不上传任何拆分上下文文件。
+- 旧的 `history_split.enabled` / `history_split.trigger_after_turns` 会被读取进配置对象以保持兼容，但不会触发拆分上传，也不会影响 `current_input_file` 的默认开启。

 相关实现：

 - 配置访问器：
  [internal/config/store_accessors.go](../internal/config/store_accessors.go)
- 历史拆分：
-  [internal/httpapi/openai/history/history_split.go](../internal/httpapi/openai/history/history_split.go)
 - 当前输入转文件：
  [internal/httpapi/openai/history/current_input_file.go](../internal/httpapi/openai/history/current_input_file.go)
+- 旧历史拆分兼容壳：
+  [internal/httpapi/openai/history/history_split.go](../internal/httpapi/openai/history/history_split.go)

-history split 触发后行为：
-
-1. 旧历史消息被切出去。
-2. 旧历史会被重新序列化成一个文本文件。
-3. 真正上传的文件名固定是 `HISTORY.txt`。
-4. 文件内容内部会使用 `IGNORE` 这层包装名来闭合 DeepSeek 官网原生文件标记。
-5. 该文件上传后，其 `file_id` 会排在 `ref_file_ids` 最前面。
-6. live prompt 只保留：
-   - system / developer
-   - 最新 user turn 起的上下文
-
-历史文件内容不是普通自由文本，而是用同一套角色标记再次序列化出的 transcript：
-
-```text
-[uploaded filename]: HISTORY.txt
-[file content end]
-
-<｜begin▁of▁sentence｜><｜User｜>...<｜Assistant｜>...<｜Tool｜>...
-
-[file name]: IGNORE
-[file content begin]
-```
-
-所以“完整上下文”在当前实现里，其实通常分散在两处：
-
- `prompt` 里的 live context
- `ref_file_ids` 指向的 history transcript file
-
-当前输入转文件启用并触发时，不会同时启用 history split，也不会上传 `HISTORY.txt`。上传文件的真实文件名是 `IGNORE.txt`，文件内容是完整 `messages` 上下文；它仍会先用 OpenAI 消息标准化和 DeepSeek 角色标记序列化，再包进 `IGNORE` 文件边界里：
+当前输入转文件启用并触发时，上传文件的真实文件名是 `IGNORE.txt`，文件内容是完整 `messages` 上下文；它仍会先用 OpenAI 消息标准化和 DeepSeek 角色标记序列化，再包进 `IGNORE` 文件边界里：

 ```text
 [uploaded filename]: IGNORE.txt
@@ -309,7 +279,7 @@ history split 触发后行为：
 - Responses `instructions` 会 prepend 为 system message
 - `tools` 会注入 system prompt
 - `attachments` / `input_file` / inline 文件会进入 `ref_file_ids`
- history split 主要在这条链路里生效
+- current input file 主要在这条链路里生效，旧 `history_split` 仅作兼容字段保留

 ### 10.2 Claude Messages

@@ -340,15 +310,15 @@ history split 触发后行为：
 - 有 tools
 - 有一个文件型 systemprompt 附件
 - 有历史 assistant tool call / tool result
- history split 已触发
+- current input file 已触发

 那么最终上下文更接近：

 ```json
 {
-  "prompt": "<｜begin▁of▁sentence｜><｜System｜>原 system / developer\n\nYou have access to these tools: ...<｜end▁of▁instructions｜><｜User｜>最新问题<｜Assistant｜>",
+  "prompt": "<｜begin▁of▁sentence｜><｜System｜>原 system / developer\n\nYou have access to these tools: ...<｜end▁of▁instructions｜><｜User｜>The current request and prior conversation context have already been provided. Answer the latest user request directly.<｜Assistant｜>",
  "ref_file_ids": [
-    "file-history-ignore",
+    "file-current-input-ignore",
    "file-systemprompt",
    "file-other-attachment"
  ],
@@ -361,7 +331,7 @@ history split 触发后行为：

 - 大部分结构化语义被压进 `prompt`
 - 文件保持文件
- 历史必要时拆文件
+- 需要时把完整上下文拆进隐藏上下文文件

 ## 12. 修改时必须同步本文档的场景

@@ -374,7 +344,8 @@ history split 触发后行为：
 - tool result 注入方式变更
 - tool prompt 模板或 tool_choice 约束变更
 - inline 文件上传 / 文件引用收集规则变更
- history split 触发条件、上传格式、`IGNORE` 包装格式变更
+- current input file 触发条件、上传格式、`IGNORE` 包装格式变更
+- 旧 `history_split` 兼容逻辑的读取、忽略或退化行为变更
 - completion payload 字段语义变更
 - Claude / Gemini 对这套统一语义的复用关系变更

--- a/internal/config/store_accessors.go
+++ b/internal/config/store_accessors.go
@@ -164,30 +164,16 @@ func (s *Store) AutoDeleteSessions() bool {
 }

 func (s *Store) HistorySplitEnabled() bool {
-	s.mu.RLock()
-	defer s.mu.RUnlock()
-	if s.cfg.HistorySplit.Enabled == nil {
-		return false
-	}
-	return *s.cfg.HistorySplit.Enabled
+	return false
 }

 func (s *Store) HistorySplitTriggerAfterTurns() int {
-	s.mu.RLock()
-	defer s.mu.RUnlock()
-	if s.cfg.HistorySplit.TriggerAfterTurns == nil || *s.cfg.HistorySplit.TriggerAfterTurns <= 0 {
-		return 1
-	}
-	return *s.cfg.HistorySplit.TriggerAfterTurns
+	return 1
 }

 func (s *Store) CurrentInputFileEnabled() bool {
 	s.mu.RLock()
 	defer s.mu.RUnlock()
-	historySplitEnabled := s.cfg.HistorySplit.Enabled != nil && *s.cfg.HistorySplit.Enabled
-	if historySplitEnabled {
-		return false
-	}
 	if s.cfg.CurrentInputFile.Enabled == nil {
 		return true
 	}
--- a/internal/config/store_accessors_test.go
+++ b/internal/config/store_accessors_test.go
@@ -3,41 +3,17 @@ package config
 import "testing"

 func TestStoreHistorySplitAccessors(t *testing.T) {
-	store := &Store{cfg: Config{}}
-	if store.HistorySplitEnabled() {
-		t.Fatal("expected history split disabled by default")
-	}
-	if got := store.HistorySplitTriggerAfterTurns(); got != 1 {
-		t.Fatalf("default history split trigger_after_turns=%d want=1", got)
-	}
-
 	enabled := true
 	turns := 3
-	store.cfg.HistorySplit = HistorySplitConfig{
+	store := &Store{cfg: Config{HistorySplit: HistorySplitConfig{
 		Enabled:           &enabled,
 		TriggerAfterTurns: &turns,
-	}
-
-	if !store.HistorySplitEnabled() {
-		t.Fatal("expected history split enabled")
-	}
-	if got := store.HistorySplitTriggerAfterTurns(); got != 3 {
-		t.Fatalf("history split trigger_after_turns=%d want=3", got)
-	}
-}
-
-func TestStoreHistorySplitDisabledConfigStaysDisabled(t *testing.T) {
-	t.Setenv("DS2API_CONFIG_JSON", `{"keys":["k1"],"history_split":{"enabled":false,"trigger_after_turns":2}}`)
-	store := LoadStore()
+	}}}
 	if store.HistorySplitEnabled() {
-		t.Fatal("expected history split disabled when config disables it")
+		t.Fatal("expected history split to stay disabled")
 	}
-	snap := store.Snapshot()
-	if snap.HistorySplit.Enabled == nil || *snap.HistorySplit.Enabled {
-		t.Fatalf("expected history_split.enabled=false, got %#v", snap.HistorySplit.Enabled)
-	}
-	if got := store.HistorySplitTriggerAfterTurns(); got != 2 {
-		t.Fatalf("history split trigger_after_turns=%d want=2", got)
+	if got := store.HistorySplitTriggerAfterTurns(); got != 1 {
+		t.Fatalf("history split trigger_after_turns=%d want=1", got)
 	}
 }

@@ -67,8 +43,8 @@ func TestStoreCurrentInputFileAccessors(t *testing.T) {

 	historyEnabled := true
 	store.cfg.HistorySplit.Enabled = &historyEnabled
-	if store.CurrentInputFileEnabled() {
-		t.Fatal("expected history split to suppress current input file mode")
+	if !store.CurrentInputFileEnabled() {
+		t.Fatal("expected history split config to not suppress current input file mode")
 	}
 }

--- a/internal/config/validation.go
+++ b/internal/config/validation.go
@@ -24,15 +24,9 @@ func ValidateConfig(c Config) error {
 	if err := ValidateAutoDeleteConfig(c.AutoDelete); err != nil {
 		return err
 	}
-	if err := ValidateHistorySplitConfig(c.HistorySplit); err != nil {
-		return err
-	}
 	if err := ValidateCurrentInputFileConfig(c.CurrentInputFile); err != nil {
 		return err
 	}
-	if c.HistorySplit.Enabled != nil && *c.HistorySplit.Enabled && c.CurrentInputFile.Enabled != nil && *c.CurrentInputFile.Enabled {
-		return fmt.Errorf("history_split.enabled and current_input_file.enabled cannot both be true")
-	}
 	if err := ValidateAccountProxyReferences(c.Accounts, c.Proxies); err != nil {
 		return err
 	}
@@ -120,15 +114,6 @@ func ValidateAutoDeleteConfig(autoDelete AutoDeleteConfig) error {
 	return ValidateAutoDeleteMode(autoDelete.Mode)
 }

-func ValidateHistorySplitConfig(historySplit HistorySplitConfig) error {
-	if historySplit.TriggerAfterTurns != nil {
-		if err := ValidateIntRange("history_split.trigger_after_turns", *historySplit.TriggerAfterTurns, 1, 1000, true); err != nil {
-			return err
-		}
-	}
-	return nil
-}
-
 func ValidateCurrentInputFileConfig(currentInputFile CurrentInputFileConfig) error {
 	if currentInputFile.MinChars != 0 {
 		return ValidateIntRange("current_input_file.min_chars", currentInputFile.MinChars, 1, 100000000, true)
--- a/internal/config/validation_test.go
+++ b/internal/config/validation_test.go
@@ -39,26 +39,11 @@ func TestValidateConfigRejectsInvalidValues(t *testing.T) {
 			cfg:  Config{AutoDelete: AutoDeleteConfig{Mode: "maybe"}},
 			want: "auto_delete.mode",
 		},
-		{
-			name: "history split",
-			cfg: Config{HistorySplit: HistorySplitConfig{
-				TriggerAfterTurns: intPtr(0),
-			}},
-			want: "history_split.trigger_after_turns",
-		},
 		{
 			name: "current input file",
 			cfg:  Config{CurrentInputFile: CurrentInputFileConfig{MinChars: -1}},
 			want: "current_input_file.min_chars",
 		},
-		{
-			name: "split modes mutually exclusive",
-			cfg: Config{
-				HistorySplit:     HistorySplitConfig{Enabled: boolPtr(true)},
-				CurrentInputFile: CurrentInputFileConfig{Enabled: boolPtr(true)},
-			},
-			want: "cannot both be true",
-		},
 	}

 	for _, tc := range tests {
@@ -79,7 +64,3 @@ func TestValidateConfigAcceptsLegacyAutoDeleteSessions(t *testing.T) {
 		t.Fatalf("expected legacy auto_delete.sessions config to remain valid, got %v", err)
 	}
 }
-
-func intPtr(v int) *int { return &v }
-
-func boolPtr(v bool) *bool { return &v }
--- a/internal/httpapi/admin/handler_settings_test.go
+++ b/internal/httpapi/admin/handler_settings_test.go
@@ -47,7 +47,7 @@ func TestGetSettingsIncludesTokenRefreshInterval(t *testing.T) {
 	}
 }

-func TestGetSettingsIncludesHistorySplitDefaults(t *testing.T) {
+func TestGetSettingsIncludesCurrentInputFileDefaults(t *testing.T) {
 	h := newAdminTestHandler(t, `{"keys":["k1"]}`)
 	req := httptest.NewRequest(http.MethodGet, "/admin/settings", nil)
 	rec := httptest.NewRecorder()
@@ -57,13 +57,6 @@ func TestGetSettingsIncludesHistorySplitDefaults(t *testing.T) {
 	}
 	var body map[string]any
 	_ = json.Unmarshal(rec.Body.Bytes(), &body)
-	historySplit, _ := body["history_split"].(map[string]any)
-	if got := boolFrom(historySplit["enabled"]); got {
-		t.Fatalf("expected history_split.enabled=false, body=%v", body)
-	}
-	if got := intFrom(historySplit["trigger_after_turns"]); got != 1 {
-		t.Fatalf("expected history_split.trigger_after_turns=1, got %d body=%v", got, body)
-	}
 	currentInputFile, _ := body["current_input_file"].(map[string]any)
 	if got := boolFrom(currentInputFile["enabled"]); !got {
 		t.Fatalf("expected current_input_file.enabled=true, body=%v", body)
@@ -190,33 +183,6 @@ func TestUpdateSettingsWithoutRuntimeSkipsMergedRuntimeValidation(t *testing.T)
 	}
 }

-func TestUpdateSettingsHistorySplit(t *testing.T) {
-	h := newAdminTestHandler(t, `{"keys":["k1"]}`)
-	payload := map[string]any{
-		"history_split": map[string]any{
-			"enabled":             true,
-			"trigger_after_turns": 3,
-		},
-	}
-	b, _ := json.Marshal(payload)
-	req := httptest.NewRequest(http.MethodPut, "/admin/settings", bytes.NewReader(b))
-	rec := httptest.NewRecorder()
-	h.updateSettings(rec, req)
-	if rec.Code != http.StatusOK {
-		t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
-	}
-	snap := h.Store.Snapshot()
-	if snap.HistorySplit.Enabled == nil || !*snap.HistorySplit.Enabled {
-		t.Fatalf("expected history_split.enabled=true, got %#v", snap.HistorySplit.Enabled)
-	}
-	if snap.HistorySplit.TriggerAfterTurns == nil || *snap.HistorySplit.TriggerAfterTurns != 3 {
-		t.Fatalf("expected history_split.trigger_after_turns=3, got %#v", snap.HistorySplit.TriggerAfterTurns)
-	}
-	if snap.CurrentInputFile.Enabled == nil || *snap.CurrentInputFile.Enabled {
-		t.Fatalf("expected history split to disable current_input_file, got %#v", snap.CurrentInputFile.Enabled)
-	}
-}
-
 func TestUpdateSettingsCurrentInputFile(t *testing.T) {
 	h := newAdminTestHandler(t, `{"keys":["k1"],"history_split":{"enabled":true,"trigger_after_turns":2}}`)
 	payload := map[string]any{
@@ -239,8 +205,11 @@ func TestUpdateSettingsCurrentInputFile(t *testing.T) {
 	if snap.CurrentInputFile.MinChars != 12345 {
 		t.Fatalf("expected current_input_file.min_chars=12345, got %#v", snap.CurrentInputFile)
 	}
-	if snap.HistorySplit.Enabled == nil || *snap.HistorySplit.Enabled {
-		t.Fatalf("expected current input file to disable history_split, got %#v", snap.HistorySplit.Enabled)
+	if !h.Store.CurrentInputFileEnabled() {
+		t.Fatal("expected current input file accessor to stay enabled")
+	}
+	if h.Store.HistorySplitEnabled() {
+		t.Fatal("expected history split accessor to stay disabled")
 	}
 }

@@ -290,7 +259,7 @@ func TestUpdateSettingsCurrentInputFilePartialUpdatePreservesMinChars(t *testing
 	}
 }

-func TestUpdateSettingsRejectsTwoSplitModesEnabled(t *testing.T) {
+func TestUpdateSettingsIgnoresHistorySplitPayload(t *testing.T) {
 	h := newAdminTestHandler(t, `{"keys":["k1"]}`)
 	payload := map[string]any{
 		"history_split": map[string]any{
@@ -306,8 +275,12 @@ func TestUpdateSettingsRejectsTwoSplitModesEnabled(t *testing.T) {
 	req := httptest.NewRequest(http.MethodPut, "/admin/settings", bytes.NewReader(b))
 	rec := httptest.NewRecorder()
 	h.updateSettings(rec, req)
-	if rec.Code != http.StatusBadRequest {
-		t.Fatalf("expected 400, got %d body=%s", rec.Code, rec.Body.String())
+	if rec.Code != http.StatusOK {
+		t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
+	}
+	snap := h.Store.Snapshot()
+	if snap.CurrentInputFile.Enabled == nil || !*snap.CurrentInputFile.Enabled {
+		t.Fatalf("expected current_input_file to remain enabled, got %#v", snap.CurrentInputFile.Enabled)
 	}
 }

--- a/internal/httpapi/admin/settings/handler_settings_parse.go
+++ b/internal/httpapi/admin/settings/handler_settings_parse.go
@@ -21,7 +21,7 @@ func boolFrom(v any) bool {
 	}
 }

-func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *config.RuntimeConfig, *config.CompatConfig, *config.ResponsesConfig, *config.EmbeddingsConfig, *config.AutoDeleteConfig, *config.HistorySplitConfig, *config.CurrentInputFileConfig, *config.ThinkingInjectionConfig, map[string]string, error) {
+func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *config.RuntimeConfig, *config.CompatConfig, *config.ResponsesConfig, *config.EmbeddingsConfig, *config.AutoDeleteConfig, *config.CurrentInputFileConfig, *config.ThinkingInjectionConfig, map[string]string, error) {
 	var (
 		adminCfg        *config.AdminConfig
 		runtimeCfg      *config.RuntimeConfig
@@ -29,7 +29,6 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		respCfg         *config.ResponsesConfig
 		embCfg          *config.EmbeddingsConfig
 		autoDeleteCfg   *config.AutoDeleteConfig
-		historySplitCfg *config.HistorySplitConfig
 		currentInputCfg *config.CurrentInputFileConfig
 		thinkingInjCfg  *config.ThinkingInjectionConfig
 		aliasMap        map[string]string
@@ -40,7 +39,7 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		if v, exists := raw["jwt_expire_hours"]; exists {
 			n := intFrom(v)
 			if err := config.ValidateIntRange("admin.jwt_expire_hours", n, 1, 720, true); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			cfg.JWTExpireHours = n
 		}
@@ -52,33 +51,33 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		if v, exists := raw["account_max_inflight"]; exists {
 			n := intFrom(v)
 			if err := config.ValidateIntRange("runtime.account_max_inflight", n, 1, 256, true); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			cfg.AccountMaxInflight = n
 		}
 		if v, exists := raw["account_max_queue"]; exists {
 			n := intFrom(v)
 			if err := config.ValidateIntRange("runtime.account_max_queue", n, 1, 200000, true); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			cfg.AccountMaxQueue = n
 		}
 		if v, exists := raw["global_max_inflight"]; exists {
 			n := intFrom(v)
 			if err := config.ValidateIntRange("runtime.global_max_inflight", n, 1, 200000, true); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			cfg.GlobalMaxInflight = n
 		}
 		if v, exists := raw["token_refresh_interval_hours"]; exists {
 			n := intFrom(v)
 			if err := config.ValidateIntRange("runtime.token_refresh_interval_hours", n, 1, 720, true); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			cfg.TokenRefreshIntervalHours = n
 		}
 		if cfg.AccountMaxInflight > 0 && cfg.GlobalMaxInflight > 0 && cfg.GlobalMaxInflight < cfg.AccountMaxInflight {
-			return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, fmt.Errorf("runtime.global_max_inflight must be >= runtime.account_max_inflight")
+			return nil, nil, nil, nil, nil, nil, nil, nil, nil, fmt.Errorf("runtime.global_max_inflight must be >= runtime.account_max_inflight")
 		}
 		runtimeCfg = cfg
 	}
@@ -101,7 +100,7 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		if v, exists := raw["store_ttl_seconds"]; exists {
 			n := intFrom(v)
 			if err := config.ValidateIntRange("responses.store_ttl_seconds", n, 30, 86400, true); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			cfg.StoreTTLSeconds = n
 		}
@@ -113,7 +112,7 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		if v, exists := raw["provider"]; exists {
 			p := strings.TrimSpace(fmt.Sprintf("%v", v))
 			if err := config.ValidateTrimmedString("embeddings.provider", p, false); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			cfg.Provider = p
 		}
@@ -139,7 +138,7 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		if v, exists := raw["mode"]; exists {
 			mode := strings.ToLower(strings.TrimSpace(fmt.Sprintf("%v", v)))
 			if err := config.ValidateAutoDeleteMode(mode); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			if mode == "" {
 				mode = "none"
@@ -152,25 +151,6 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		autoDeleteCfg = cfg
 	}

-	if raw, ok := req["history_split"].(map[string]any); ok {
-		cfg := &config.HistorySplitConfig{}
-		if v, exists := raw["enabled"]; exists {
-			enabled := boolFrom(v)
-			cfg.Enabled = &enabled
-		}
-		if v, exists := raw["trigger_after_turns"]; exists {
-			n := intFrom(v)
-			if err := config.ValidateIntRange("history_split.trigger_after_turns", n, 1, 1000, true); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
-			}
-			cfg.TriggerAfterTurns = &n
-		}
-		if err := config.ValidateHistorySplitConfig(*cfg); err != nil {
-			return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
-		}
-		historySplitCfg = cfg
-	}
-
 	if raw, ok := req["current_input_file"].(map[string]any); ok {
 		cfg := &config.CurrentInputFileConfig{}
 		if v, exists := raw["enabled"]; exists {
@@ -180,18 +160,15 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		if v, exists := raw["min_chars"]; exists {
 			n := intFrom(v)
 			if err := config.ValidateIntRange("current_input_file.min_chars", n, 0, 100000000, true); err != nil {
-				return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+				return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 			}
 			cfg.MinChars = n
 		}
 		if err := config.ValidateCurrentInputFileConfig(*cfg); err != nil {
-			return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, err
+			return nil, nil, nil, nil, nil, nil, nil, nil, nil, err
 		}
 		currentInputCfg = cfg
 	}
-	if boolPtrValue(historySplitCfgEnabled(historySplitCfg)) && boolPtrValue(currentInputCfgEnabled(currentInputCfg)) {
-		return nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, fmt.Errorf("history_split.enabled and current_input_file.enabled cannot both be true")
-	}

 	if raw, ok := req["thinking_injection"].(map[string]any); ok {
 		cfg := &config.ThinkingInjectionConfig{}
@@ -205,23 +182,5 @@ func parseSettingsUpdateRequest(req map[string]any) (*config.AdminConfig, *confi
 		thinkingInjCfg = cfg
 	}

-	return adminCfg, runtimeCfg, compatCfg, respCfg, embCfg, autoDeleteCfg, historySplitCfg, currentInputCfg, thinkingInjCfg, aliasMap, nil
-}
-
-func historySplitCfgEnabled(cfg *config.HistorySplitConfig) *bool {
-	if cfg == nil {
-		return nil
-	}
-	return cfg.Enabled
-}
-
-func currentInputCfgEnabled(cfg *config.CurrentInputFileConfig) *bool {
-	if cfg == nil {
-		return nil
-	}
-	return cfg.Enabled
-}
-
-func boolPtrValue(v *bool) bool {
-	return v != nil && *v
+	return adminCfg, runtimeCfg, compatCfg, respCfg, embCfg, autoDeleteCfg, currentInputCfg, thinkingInjCfg, aliasMap, nil
 }
--- a/internal/httpapi/admin/settings/handler_settings_read.go
+++ b/internal/httpapi/admin/settings/handler_settings_read.go
@@ -31,10 +31,6 @@ func (h *Handler) getSettings(w http.ResponseWriter, _ *http.Request) {
 		"responses":   snap.Responses,
 		"embeddings":  snap.Embeddings,
 		"auto_delete": snap.AutoDelete,
-		"history_split": map[string]any{
-			"enabled":             h.Store.HistorySplitEnabled(),
-			"trigger_after_turns": h.Store.HistorySplitTriggerAfterTurns(),
-		},
 		"current_input_file": map[string]any{
 			"enabled":   h.Store.CurrentInputFileEnabled(),
 			"min_chars": h.Store.CurrentInputFileMinChars(),
--- a/internal/httpapi/admin/settings/handler_settings_write.go
+++ b/internal/httpapi/admin/settings/handler_settings_write.go
@@ -17,7 +17,7 @@ func (h *Handler) updateSettings(w http.ResponseWriter, r *http.Request) {
 		return
 	}

-	adminCfg, runtimeCfg, compatCfg, responsesCfg, embeddingsCfg, autoDeleteCfg, historySplitCfg, currentInputCfg, thinkingInjCfg, aliasMap, err := parseSettingsUpdateRequest(req)
+	adminCfg, runtimeCfg, compatCfg, responsesCfg, embeddingsCfg, autoDeleteCfg, currentInputCfg, thinkingInjCfg, aliasMap, err := parseSettingsUpdateRequest(req)
 	if err != nil {
 		writeJSON(w, http.StatusBadRequest, map[string]any{"detail": err.Error()})
 		return
@@ -71,26 +71,10 @@ func (h *Handler) updateSettings(w http.ResponseWriter, r *http.Request) {
 			c.AutoDelete.Mode = autoDeleteCfg.Mode
 			c.AutoDelete.Sessions = autoDeleteCfg.Sessions
 		}
-		if historySplitCfg != nil {
-			if historySplitCfg.Enabled != nil {
-				c.HistorySplit.Enabled = historySplitCfg.Enabled
-				if *historySplitCfg.Enabled {
-					disabled := false
-					c.CurrentInputFile.Enabled = &disabled
-				}
-			}
-			if historySplitCfg.TriggerAfterTurns != nil {
-				c.HistorySplit.TriggerAfterTurns = historySplitCfg.TriggerAfterTurns
-			}
-		}
 		if currentInputCfg != nil {
 			if currentInputEnabledSet {
 				c.CurrentInputFile.Enabled = currentInputCfg.Enabled
 			}
-			if currentInputEnabledSet && currentInputCfg.Enabled != nil && *currentInputCfg.Enabled {
-				disabled := false
-				c.HistorySplit.Enabled = &disabled
-			}
 			if currentInputMinCharsSet {
 				c.CurrentInputFile.MinChars = currentInputCfg.MinChars
 			}
--- a/internal/httpapi/openai/chat/chat_history_test.go
+++ b/internal/httpapi/openai/chat/chat_history_test.go
@@ -272,14 +272,13 @@ func TestChatCompletionsSkipsHistoryWhenDisabled(t *testing.T) {
 	}
 }

-func TestChatCompletionsHistorySplitPersistsHistoryText(t *testing.T) {
+func TestChatCompletionsCurrentInputFilePersistsNeutralPrompt(t *testing.T) {
 	historyStore := newTestChatHistoryStore(t)
 	ds := &inlineUploadDSStub{}
 	h := &Handler{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		Auth:        streamStatusAuthStub{},
 		DS:          ds,
@@ -308,19 +307,19 @@ func TestChatCompletionsHistorySplitPersistsHistoryText(t *testing.T) {
 	if err != nil {
 		t.Fatalf("expected detail item, got %v", err)
 	}
-	if full.HistoryText == "" {
-		t.Fatalf("expected history text to be persisted")
-	}
-	if !strings.Contains(full.HistoryText, "first user turn") || !strings.Contains(full.HistoryText, "tool result") {
-		t.Fatalf("expected earlier turns in history text, got %q", full.HistoryText)
-	}
-	if strings.Contains(full.HistoryText, "latest user turn") {
-		t.Fatalf("expected latest turn to stay out of persisted history text, got %q", full.HistoryText)
+	if full.HistoryText != "" {
+		t.Fatalf("expected current input file flow to leave history text empty, got %q", full.HistoryText)
 	}
 	if len(ds.uploadCalls) != 1 {
-		t.Fatalf("expected history upload to happen, got %d", len(ds.uploadCalls))
+		t.Fatalf("expected current input upload to happen, got %d", len(ds.uploadCalls))
 	}
-	if full.HistoryText != string(ds.uploadCalls[0].Data) {
-		t.Fatalf("expected persisted history text to match uploaded HISTORY.txt contents")
+	if ds.uploadCalls[0].Filename != "IGNORE.txt" {
+		t.Fatalf("expected IGNORE.txt upload, got %q", ds.uploadCalls[0].Filename)
+	}
+	if len(full.Messages) != 1 {
+		t.Fatalf("expected neutral prompt to be the only persisted message, got %#v", full.Messages)
+	}
+	if !strings.Contains(full.Messages[0].Content, "Answer the latest user request directly.") {
+		t.Fatalf("expected neutral prompt to be persisted, got %#v", full.Messages[0])
 	}
 }
--- a/internal/httpapi/openai/chat/handler.go
+++ b/internal/httpapi/openai/chat/handler.go
@@ -42,20 +42,17 @@ func (h *Handler) compatStripReferenceMarkers() bool {
 	return shared.CompatStripReferenceMarkers(h.Store)
 }

-func (h *Handler) applyHistorySplit(ctx context.Context, a *auth.RequestAuth, stdReq promptcompat.StandardRequest) (promptcompat.StandardRequest, error) {
+func (h *Handler) applyCurrentInputFile(ctx context.Context, a *auth.RequestAuth, stdReq promptcompat.StandardRequest) (promptcompat.StandardRequest, error) {
 	if h == nil {
 		return stdReq, nil
 	}
 	stdReq = shared.ApplyThinkingInjection(h.Store, stdReq)
 	svc := history.Service{Store: h.Store, DS: h.DS}
 	out, err := svc.ApplyCurrentInputFile(ctx, a, stdReq)
-	if err != nil {
-		return stdReq, err
+	if err != nil || out.CurrentInputFileApplied {
+		return out, err
 	}
-	if out.CurrentInputFileApplied {
-		return out, nil
-	}
-	return svc.Apply(ctx, a, out)
+	return out, nil
 }

 func (h *Handler) preprocessInlineFileInputs(ctx context.Context, a *auth.RequestAuth, req map[string]any) error {
@@ -91,7 +88,7 @@ func writeOpenAIInlineFileError(w http.ResponseWriter, err error) {
 	files.WriteInlineFileError(w, err)
 }

-func mapHistorySplitError(err error) (int, string) {
+func mapCurrentInputFileError(err error) (int, string) {
 	return history.MapError(err)
 }

--- a/internal/httpapi/openai/chat/handler_chat.go
+++ b/internal/httpapi/openai/chat/handler_chat.go
@@ -68,9 +68,9 @@ func (h *Handler) ChatCompletions(w http.ResponseWriter, r *http.Request) {
 		writeOpenAIError(w, http.StatusBadRequest, err.Error())
 		return
 	}
-	stdReq, err = h.applyHistorySplit(r.Context(), a, stdReq)
+	stdReq, err = h.applyCurrentInputFile(r.Context(), a, stdReq)
 	if err != nil {
-		status, message := mapHistorySplitError(err)
+		status, message := mapCurrentInputFileError(err)
 		writeOpenAIError(w, status, message)
 		return
 	}
--- a/internal/httpapi/openai/chat/vercel_prepare_test.go
+++ b/internal/httpapi/openai/chat/vercel_prepare_test.go
@@ -87,7 +87,7 @@ func TestStreamLeaseTTL(t *testing.T) {
 	}
 }

-func TestHandleVercelStreamPrepareAppliesHistorySplit(t *testing.T) {
+func TestHandleVercelStreamPrepareAppliesCurrentInputFile(t *testing.T) {
 	t.Setenv("VERCEL", "1")
 	t.Setenv("DS2API_VERCEL_INTERNAL_SECRET", "stream-secret")

@@ -95,8 +95,7 @@ func TestHandleVercelStreamPrepareAppliesHistorySplit(t *testing.T) {
 	h := &Handler{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		Auth: streamStatusAuthStub{},
 		DS:   ds,
@@ -119,7 +118,7 @@ func TestHandleVercelStreamPrepareAppliesHistorySplit(t *testing.T) {
 		t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
 	}
 	if len(ds.uploadCalls) != 1 {
-		t.Fatalf("expected 1 history upload, got %d", len(ds.uploadCalls))
+		t.Fatalf("expected 1 current input upload, got %d", len(ds.uploadCalls))
 	}

 	var body map[string]any
@@ -131,11 +130,11 @@ func TestHandleVercelStreamPrepareAppliesHistorySplit(t *testing.T) {
 		t.Fatalf("expected payload object, got %#v", body["payload"])
 	}
 	promptText, _ := payload["prompt"].(string)
-	if !strings.Contains(promptText, "latest user turn") {
-		t.Fatalf("expected latest user turn in prompt, got %s", promptText)
+	if !strings.Contains(promptText, "Answer the latest user request directly.") {
+		t.Fatalf("expected neutral prompt, got %s", promptText)
 	}
-	if strings.Contains(promptText, "first user turn") {
-		t.Fatalf("expected historical turns removed from prompt, got %s", promptText)
+	if strings.Contains(promptText, "first user turn") || strings.Contains(promptText, "latest user turn") {
+		t.Fatalf("expected original turns hidden from prompt, got %s", promptText)
 	}
 	refIDs, _ := payload["ref_file_ids"].([]any)
 	if len(refIDs) == 0 || refIDs[0] != "file-inline-1" {
@@ -143,7 +142,7 @@ func TestHandleVercelStreamPrepareAppliesHistorySplit(t *testing.T) {
 	}
 }

-func TestHandleVercelStreamPrepareMapsHistorySplitManagedAuthFailureTo401(t *testing.T) {
+func TestHandleVercelStreamPrepareMapsCurrentInputFileManagedAuthFailureTo401(t *testing.T) {
 	t.Setenv("VERCEL", "1")
 	t.Setenv("DS2API_VERCEL_INTERNAL_SECRET", "stream-secret")

@@ -153,8 +152,7 @@ func TestHandleVercelStreamPrepareMapsHistorySplitManagedAuthFailureTo401(t *tes
 	h := &Handler{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		Auth: streamStatusManagedAuthStub{},
 		DS:   ds,
--- a/internal/httpapi/openai/chat/vercel_stream.go
+++ b/internal/httpapi/openai/chat/vercel_stream.go
@@ -69,9 +69,9 @@ func (h *Handler) handleVercelStreamPrepare(w http.ResponseWriter, r *http.Reque
 		writeOpenAIError(w, http.StatusBadRequest, "stream must be true")
 		return
 	}
-	stdReq, err = h.applyHistorySplit(r.Context(), a, stdReq)
+	stdReq, err = h.applyCurrentInputFile(r.Context(), a, stdReq)
 	if err != nil {
-		status, message := mapHistorySplitError(err)
+		status, message := mapCurrentInputFileError(err)
 		writeOpenAIError(w, status, message)
 		return
 	}
--- a/internal/httpapi/openai/history/current_input_file.go
+++ b/internal/httpapi/openai/history/current_input_file.go
@@ -28,8 +28,7 @@ func (s Service) ApplyCurrentInputFile(ctx context.Context, a *auth.RequestAuth,
 	if index < 0 {
 		return stdReq, nil
 	}
-	historySplitReached := s.Store.HistorySplitEnabled() && wouldSplitHistory(stdReq.Messages, s.Store.HistorySplitTriggerAfterTurns())
-	if len([]rune(text)) < threshold && !historySplitReached {
+	if len([]rune(text)) < threshold {
 		return stdReq, nil
 	}
 	fileText := promptcompat.BuildOpenAICurrentInputContextTranscript(stdReq.Messages)
@@ -84,11 +83,6 @@ func latestUserInputForFile(messages []any) (int, string) {
 	return -1, ""
 }

-func wouldSplitHistory(messages []any, triggerAfterTurns int) bool {
-	_, historyMessages := SplitOpenAIHistoryMessages(messages, triggerAfterTurns)
-	return len(historyMessages) > 0
-}
-
 func currentInputFilePrompt() string {
 	return "The current request and prior conversation context have already been provided. Answer the latest user request directly."
 }
--- a/internal/httpapi/openai/history/history_split.go
+++ b/internal/httpapi/openai/history/history_split.go
@@ -2,60 +2,21 @@ package history

 import (
 	"context"
-	"errors"
-	"fmt"
 	"strings"

 	"ds2api/internal/auth"
-	dsclient "ds2api/internal/deepseek/client"
 	"ds2api/internal/httpapi/openai/shared"
 	"ds2api/internal/promptcompat"
 )

-const (
-	historySplitFilename    = "HISTORY.txt"
-	historySplitContentType = "text/plain; charset=utf-8"
-	historySplitPurpose     = "assistants"
-)
-
 type Service struct {
 	Store shared.ConfigReader
 	DS    shared.DeepSeekCaller
 }

+// Apply is retained for legacy compatibility only. The active split path is
+// current input file handling in ApplyCurrentInputFile.
 func (s Service) Apply(ctx context.Context, a *auth.RequestAuth, stdReq promptcompat.StandardRequest) (promptcompat.StandardRequest, error) {
-	if s.DS == nil || s.Store == nil || a == nil || !s.Store.HistorySplitEnabled() {
-		return stdReq, nil
-	}
-
-	promptMessages, historyMessages := SplitOpenAIHistoryMessages(stdReq.Messages, s.Store.HistorySplitTriggerAfterTurns())
-	if len(historyMessages) == 0 {
-		return stdReq, nil
-	}
-
-	historyText := promptcompat.BuildOpenAIHistoryTranscript(historyMessages)
-	if strings.TrimSpace(historyText) == "" {
-		return stdReq, errors.New("history split produced empty transcript")
-	}
-
-	result, err := s.DS.UploadFile(ctx, a, dsclient.UploadFileRequest{
-		Filename:    historySplitFilename,
-		ContentType: historySplitContentType,
-		Purpose:     historySplitPurpose,
-		Data:        []byte(historyText),
-	}, 3)
-	if err != nil {
-		return stdReq, fmt.Errorf("upload history file: %w", err)
-	}
-	fileID := strings.TrimSpace(result.ID)
-	if fileID == "" {
-		return stdReq, errors.New("upload history file returned empty file id")
-	}
-
-	stdReq.Messages = promptMessages
-	stdReq.HistoryText = historyText
-	stdReq.RefFileIDs = prependUniqueRefFileID(stdReq.RefFileIDs, fileID)
-	stdReq.FinalPrompt, stdReq.ToolNames = promptcompat.BuildOpenAIPrompt(promptMessages, stdReq.ToolsRaw, "", stdReq.ToolChoice, stdReq.Thinking)
 	return stdReq, nil
 }

--- a/internal/httpapi/openai/history_split_test.go
+++ b/internal/httpapi/openai/history_split_test.go
@@ -60,9 +60,9 @@ func (streamStatusManagedAuthStub) DetermineCaller(_ *http.Request) (*auth.Reque

 func (streamStatusManagedAuthStub) Release(_ *auth.RequestAuth) {}

-func TestBuildOpenAIHistoryTranscriptUsesInjectedFileWrapper(t *testing.T) {
+func TestBuildOpenAICurrentInputContextTranscriptUsesInjectedFileWrapper(t *testing.T) {
 	_, historyMessages := splitOpenAIHistoryMessages(historySplitTestMessages(), 1)
-	transcript := buildOpenAIHistoryTranscript(historyMessages)
+	transcript := buildOpenAICurrentInputContextTranscript(historyMessages)

 	if !strings.HasPrefix(transcript, "[file content end]\n\n") {
 		t.Fatalf("expected injected file wrapper prefix, got %q", transcript)
@@ -107,7 +107,7 @@ func TestSplitOpenAIHistoryMessagesUsesLatestUserTurn(t *testing.T) {
 		t.Fatalf("expected middle user turn to be moved into history, got %s", promptText)
 	}

-	historyText := buildOpenAIHistoryTranscript(historyMessages)
+	historyText := buildOpenAICurrentInputContextTranscript(historyMessages)
 	if !strings.Contains(historyText, "middle user turn") {
 		t.Fatalf("expected middle user turn in split history, got %s", historyText)
 	}
@@ -116,13 +116,13 @@ func TestSplitOpenAIHistoryMessagesUsesLatestUserTurn(t *testing.T) {
 	}
 }

-func TestApplyHistorySplitSkipsFirstTurn(t *testing.T) {
+func TestApplyCurrentInputFileSkipsShortInputWhenThresholdNotReached(t *testing.T) {
 	ds := &inlineUploadDSStub{}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
+			currentInputMin:     10,
 		},
 		DS: ds,
 	}
@@ -137,9 +137,9 @@ func TestApplyHistorySplitSkipsFirstTurn(t *testing.T) {
 		t.Fatalf("normalize failed: %v", err)
 	}

-	out, err := h.applyHistorySplit(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
+	out, err := h.applyCurrentInputFile(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
 	if err != nil {
-		t.Fatalf("apply history split failed: %v", err)
+		t.Fatalf("apply current input file failed: %v", err)
 	}
 	if len(ds.uploadCalls) != 0 {
 		t.Fatalf("expected no upload on first turn, got %d", len(ds.uploadCalls))
@@ -153,10 +153,8 @@ func TestApplyThinkingInjectionAppendsLatestUserPrompt(t *testing.T) {
 	ds := &inlineUploadDSStub{}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
-			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
-			thinkingInjection:   boolPtr(true),
+			wideInput:         true,
+			thinkingInjection: boolPtr(true),
 		},
 		DS: ds,
 	}
@@ -171,7 +169,7 @@ func TestApplyThinkingInjectionAppendsLatestUserPrompt(t *testing.T) {
 		t.Fatalf("normalize failed: %v", err)
 	}

-	out, err := h.applyHistorySplit(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
+	out, err := h.applyCurrentInputFile(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
 	if err != nil {
 		t.Fatalf("apply thinking injection failed: %v", err)
 	}
@@ -204,7 +202,7 @@ func TestApplyThinkingInjectionUsesCustomPrompt(t *testing.T) {
 		t.Fatalf("normalize failed: %v", err)
 	}

-	out, err := h.applyHistorySplit(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
+	out, err := h.applyCurrentInputFile(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
 	if err != nil {
 		t.Fatalf("apply thinking injection failed: %v", err)
 	}
@@ -213,12 +211,11 @@ func TestApplyThinkingInjectionUsesCustomPrompt(t *testing.T) {
 	}
 }

-func TestApplyHistorySplitDirectPassThroughWhenBothSplitsDisabled(t *testing.T) {
+func TestApplyCurrentInputFileDisabledPassThrough(t *testing.T) {
 	ds := &inlineUploadDSStub{}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: false,
 			currentInputEnabled: false,
 		},
 		DS: ds,
@@ -232,9 +229,9 @@ func TestApplyHistorySplitDirectPassThroughWhenBothSplitsDisabled(t *testing.T)
 		t.Fatalf("normalize failed: %v", err)
 	}

-	out, err := h.applyHistorySplit(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
+	out, err := h.applyCurrentInputFile(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
 	if err != nil {
-		t.Fatalf("apply history split failed: %v", err)
+		t.Fatalf("apply current input file failed: %v", err)
 	}
 	if len(ds.uploadCalls) != 0 {
 		t.Fatalf("expected no uploads when both split modes are disabled, got %d", len(ds.uploadCalls))
@@ -252,8 +249,6 @@ func TestApplyCurrentInputFileUploadsFirstTurnWithInjectedWrapper(t *testing.T)
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
 			currentInputEnabled: true,
 			currentInputMin:     10,
 			thinkingInjection:   boolPtr(true),
@@ -271,7 +266,7 @@ func TestApplyCurrentInputFileUploadsFirstTurnWithInjectedWrapper(t *testing.T)
 		t.Fatalf("normalize failed: %v", err)
 	}

-	out, err := h.applyHistorySplit(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
+	out, err := h.applyCurrentInputFile(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
 	if err != nil {
 		t.Fatalf("apply current input file failed: %v", err)
 	}
@@ -309,15 +304,13 @@ func TestApplyCurrentInputFileUploadsFirstTurnWithInjectedWrapper(t *testing.T)
 	}
 }

-func TestApplyCurrentInputFileReplacesHistorySplitWithFullContextFile(t *testing.T) {
+func TestApplyCurrentInputFileUploadsFullContextFile(t *testing.T) {
 	ds := &inlineUploadDSStub{}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
 			currentInputEnabled: true,
-			currentInputMin:     1000,
+			currentInputMin:     0,
 			thinkingInjection:   boolPtr(true),
 		},
 		DS: ds,
@@ -331,12 +324,12 @@ func TestApplyCurrentInputFileReplacesHistorySplitWithFullContextFile(t *testing
 		t.Fatalf("normalize failed: %v", err)
 	}

-	out, err := h.applyHistorySplit(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
+	out, err := h.applyCurrentInputFile(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
 	if err != nil {
 		t.Fatalf("apply current input file failed: %v", err)
 	}
 	if !out.CurrentInputFileApplied {
-		t.Fatalf("expected current input file to replace history split")
+		t.Fatalf("expected current input file to apply")
 	}
 	if len(ds.uploadCalls) != 1 {
 		t.Fatalf("expected one current input upload, got %d", len(ds.uploadCalls))
@@ -351,9 +344,6 @@ func TestApplyCurrentInputFileReplacesHistorySplitWithFullContextFile(t *testing
 			t.Fatalf("expected full context file to contain %q, got %q", want, uploadedText)
 		}
 	}
-	if out.HistoryText != "" {
-		t.Fatalf("expected no HISTORY transcript when current input file replaces split, got %q", out.HistoryText)
-	}
 	if strings.Contains(out.FinalPrompt, "first user turn") || strings.Contains(out.FinalPrompt, "latest user turn") || strings.Contains(out.FinalPrompt, "CURRENT_USER_INPUT.txt") || strings.Contains(out.FinalPrompt, "IGNORE.txt") || strings.Contains(out.FinalPrompt, "Read that file") {
 		t.Fatalf("expected live prompt to use only a neutral continuation instruction, got %s", out.FinalPrompt)
 	}
@@ -362,13 +352,12 @@ func TestApplyCurrentInputFileReplacesHistorySplitWithFullContextFile(t *testing
 	}
 }

-func TestApplyHistorySplitCarriesHistoryText(t *testing.T) {
+func TestApplyCurrentInputFileLeavesHistoryTextEmpty(t *testing.T) {
 	ds := &inlineUploadDSStub{}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		DS: ds,
 	}
@@ -381,25 +370,24 @@ func TestApplyHistorySplitCarriesHistoryText(t *testing.T) {
 		t.Fatalf("normalize failed: %v", err)
 	}

-	out, err := h.applyHistorySplit(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
+	out, err := h.applyCurrentInputFile(context.Background(), &auth.RequestAuth{DeepSeekToken: "token"}, stdReq)
 	if err != nil {
-		t.Fatalf("apply history split failed: %v", err)
+		t.Fatalf("apply current input file failed: %v", err)
 	}
 	if len(ds.uploadCalls) != 1 {
 		t.Fatalf("expected 1 upload call, got %d", len(ds.uploadCalls))
 	}
-	if out.HistoryText != string(ds.uploadCalls[0].Data) {
-		t.Fatalf("expected history text to be preserved on normalized request")
+	if out.HistoryText != "" {
+		t.Fatalf("expected current input file flow to leave history text empty, got %q", out.HistoryText)
 	}
 }

-func TestChatCompletionsHistorySplitUploadsHistoryFileAndKeepsLatestPrompt(t *testing.T) {
+func TestChatCompletionsCurrentInputFileUploadsContextAndKeepsNeutralPrompt(t *testing.T) {
 	ds := &inlineUploadDSStub{}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		Auth: streamStatusAuthStub{},
 		DS:   ds,
@@ -423,7 +411,7 @@ func TestChatCompletionsHistorySplitUploadsHistoryFileAndKeepsLatestPrompt(t *te
 		t.Fatalf("expected 1 upload call, got %d", len(ds.uploadCalls))
 	}
 	upload := ds.uploadCalls[0]
-	if upload.Filename != "HISTORY.txt" {
+	if upload.Filename != "IGNORE.txt" {
 		t.Fatalf("unexpected upload filename: %q", upload.Filename)
 	}
 	if upload.Purpose != "assistants" {
@@ -433,32 +421,31 @@ func TestChatCompletionsHistorySplitUploadsHistoryFileAndKeepsLatestPrompt(t *te
 	if !strings.Contains(historyText, "[file content end]") || !strings.Contains(historyText, "[file name]: IGNORE") {
 		t.Fatalf("expected injected IGNORE wrapper, got %s", historyText)
 	}
-	if strings.Contains(historyText, "latest user turn") {
-		t.Fatalf("expected latest turn to remain live, got %s", historyText)
+	if !strings.Contains(historyText, "latest user turn") {
+		t.Fatalf("expected full context to include latest turn, got %s", historyText)
 	}
 	if ds.completionReq == nil {
 		t.Fatal("expected completion payload to be captured")
 	}
 	promptText, _ := ds.completionReq["prompt"].(string)
-	if !strings.Contains(promptText, "latest user turn") {
-		t.Fatalf("expected latest turn in completion prompt, got %s", promptText)
+	if !strings.Contains(promptText, "Answer the latest user request directly.") {
+		t.Fatalf("expected neutral completion prompt, got %s", promptText)
 	}
-	if strings.Contains(promptText, "first user turn") {
-		t.Fatalf("expected historical turns removed from completion prompt, got %s", promptText)
+	if strings.Contains(promptText, "first user turn") || strings.Contains(promptText, "latest user turn") {
+		t.Fatalf("expected prompt to hide original turns, got %s", promptText)
 	}
 	refIDs, _ := ds.completionReq["ref_file_ids"].([]any)
 	if len(refIDs) == 0 || refIDs[0] != "file-inline-1" {
-		t.Fatalf("expected uploaded history file to be first ref_file_id, got %#v", ds.completionReq["ref_file_ids"])
+		t.Fatalf("expected uploaded current input file to be first ref_file_id, got %#v", ds.completionReq["ref_file_ids"])
 	}
 }

-func TestResponsesHistorySplitUploadsHistoryAndKeepsLatestPrompt(t *testing.T) {
+func TestResponsesCurrentInputFileUploadsContextAndKeepsNeutralPrompt(t *testing.T) {
 	ds := &inlineUploadDSStub{}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		Auth: streamStatusAuthStub{},
 		DS:   ds,
@@ -487,23 +474,22 @@ func TestResponsesHistorySplitUploadsHistoryAndKeepsLatestPrompt(t *testing.T) {
 		t.Fatal("expected completion payload to be captured")
 	}
 	promptText, _ := ds.completionReq["prompt"].(string)
-	if !strings.Contains(promptText, "latest user turn") {
-		t.Fatalf("expected latest turn in completion prompt, got %s", promptText)
+	if !strings.Contains(promptText, "Answer the latest user request directly.") {
+		t.Fatalf("expected neutral completion prompt, got %s", promptText)
 	}
-	if strings.Contains(promptText, "first user turn") {
-		t.Fatalf("expected historical turns removed from completion prompt, got %s", promptText)
+	if strings.Contains(promptText, "first user turn") || strings.Contains(promptText, "latest user turn") {
+		t.Fatalf("expected prompt to hide original turns, got %s", promptText)
 	}
 }

-func TestChatCompletionsHistorySplitMapsManagedAuthFailureTo401(t *testing.T) {
+func TestChatCompletionsCurrentInputFileMapsManagedAuthFailureTo401(t *testing.T) {
 	ds := &inlineUploadDSStub{
 		uploadErr: &dsclient.RequestFailure{Op: "upload file", Kind: dsclient.FailureManagedUnauthorized, Message: "expired token"},
 	}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		Auth: streamStatusManagedAuthStub{},
 		DS:   ds,
@@ -528,15 +514,14 @@ func TestChatCompletionsHistorySplitMapsManagedAuthFailureTo401(t *testing.T) {
 	}
 }

-func TestResponsesHistorySplitMapsDirectAuthFailureTo401(t *testing.T) {
+func TestResponsesCurrentInputFileMapsDirectAuthFailureTo401(t *testing.T) {
 	ds := &inlineUploadDSStub{
 		uploadErr: &dsclient.RequestFailure{Op: "upload file", Kind: dsclient.FailureDirectUnauthorized, Message: "invalid token"},
 	}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		Auth: streamStatusAuthStub{},
 		DS:   ds,
@@ -563,13 +548,12 @@ func TestResponsesHistorySplitMapsDirectAuthFailureTo401(t *testing.T) {
 	}
 }

-func TestChatCompletionsHistorySplitUploadFailureReturnsInternalServerError(t *testing.T) {
+func TestChatCompletionsCurrentInputFileUploadFailureReturnsInternalServerError(t *testing.T) {
 	ds := &inlineUploadDSStub{uploadErr: errors.New("boom")}
 	h := &openAITestSurface{
 		Store: mockOpenAIConfig{
 			wideInput:           true,
-			historySplitEnabled: true,
-			historySplitTurns:   1,
+			currentInputEnabled: true,
 		},
 		Auth: streamStatusAuthStub{},
 		DS:   ds,
@@ -591,7 +575,7 @@ func TestChatCompletionsHistorySplitUploadFailureReturnsInternalServerError(t *t
 	}
 }

-func TestHistorySplitWorksAcrossAutoDeleteModes(t *testing.T) {
+func TestCurrentInputFileWorksAcrossAutoDeleteModes(t *testing.T) {
 	for _, mode := range []string{"none", "single", "all"} {
 		t.Run(mode, func(t *testing.T) {
 			ds := &inlineUploadDSStub{}
@@ -599,8 +583,7 @@ func TestHistorySplitWorksAcrossAutoDeleteModes(t *testing.T) {
 				Store: mockOpenAIConfig{
 					wideInput:           true,
 					autoDeleteMode:      mode,
-					historySplitEnabled: true,
-					historySplitTurns:   1,
+					currentInputEnabled: true,
 				},
 				Auth: streamStatusAuthStub{},
 				DS:   ds,
@@ -621,13 +604,13 @@ func TestHistorySplitWorksAcrossAutoDeleteModes(t *testing.T) {
 				t.Fatalf("expected 200, got %d body=%s", rec.Code, rec.Body.String())
 			}
 			if len(ds.uploadCalls) != 1 {
-				t.Fatalf("expected history split upload for mode=%s, got %d", mode, len(ds.uploadCalls))
+				t.Fatalf("expected current input upload for mode=%s, got %d", mode, len(ds.uploadCalls))
 			}
 			if ds.completionReq == nil {
 				t.Fatalf("expected completion payload for mode=%s", mode)
 			}
 			promptText, _ := ds.completionReq["prompt"].(string)
-			if !strings.Contains(promptText, "latest user turn") || strings.Contains(promptText, "first user turn") {
+			if !strings.Contains(promptText, "Answer the latest user request directly.") || strings.Contains(promptText, "first user turn") || strings.Contains(promptText, "latest user turn") {
 				t.Fatalf("unexpected prompt for mode=%s: %s", mode, promptText)
 			}
 		})
--- a/internal/httpapi/openai/responses/handler.go
+++ b/internal/httpapi/openai/responses/handler.go
@@ -36,20 +36,17 @@ func (h *Handler) compatStripReferenceMarkers() bool {
 	return shared.CompatStripReferenceMarkers(h.Store)
 }

-func (h *Handler) applyHistorySplit(ctx context.Context, a *auth.RequestAuth, stdReq promptcompat.StandardRequest) (promptcompat.StandardRequest, error) {
+func (h *Handler) applyCurrentInputFile(ctx context.Context, a *auth.RequestAuth, stdReq promptcompat.StandardRequest) (promptcompat.StandardRequest, error) {
 	if h == nil {
 		return stdReq, nil
 	}
 	stdReq = shared.ApplyThinkingInjection(h.Store, stdReq)
 	svc := history.Service{Store: h.Store, DS: h.DS}
 	out, err := svc.ApplyCurrentInputFile(ctx, a, stdReq)
-	if err != nil {
-		return stdReq, err
+	if err != nil || out.CurrentInputFileApplied {
+		return out, err
 	}
-	if out.CurrentInputFileApplied {
-		return out, nil
-	}
-	return svc.Apply(ctx, a, out)
+	return out, nil
 }

 func (h *Handler) preprocessInlineFileInputs(ctx context.Context, a *auth.RequestAuth, req map[string]any) error {
@@ -89,7 +86,7 @@ func writeOpenAIInlineFileError(w http.ResponseWriter, err error) {
 	files.WriteInlineFileError(w, err)
 }

-func mapHistorySplitError(err error) (int, string) {
+func mapCurrentInputFileError(err error) (int, string) {
 	return history.MapError(err)
 }

--- a/internal/httpapi/openai/responses/responses_handler.go
+++ b/internal/httpapi/openai/responses/responses_handler.go
@@ -85,9 +85,9 @@ func (h *Handler) Responses(w http.ResponseWriter, r *http.Request) {
 		writeOpenAIError(w, http.StatusBadRequest, err.Error())
 		return
 	}
-	stdReq, err = h.applyHistorySplit(r.Context(), a, stdReq)
+	stdReq, err = h.applyCurrentInputFile(r.Context(), a, stdReq)
 	if err != nil {
-		status, message := mapHistorySplitError(err)
+		status, message := mapCurrentInputFileError(err)
 		writeOpenAIError(w, status, message)
 		return
 	}
--- a/internal/httpapi/openai/test_bridge_test.go
+++ b/internal/httpapi/openai/test_bridge_test.go
@@ -83,17 +83,14 @@ func (h *openAITestSurface) ChatCompletions(w http.ResponseWriter, r *http.Reque
 	h.chatHandler().ChatCompletions(w, r)
 }

-func (h *openAITestSurface) applyHistorySplit(ctx context.Context, a *auth.RequestAuth, stdReq promptcompat.StandardRequest) (promptcompat.StandardRequest, error) {
+func (h *openAITestSurface) applyCurrentInputFile(ctx context.Context, a *auth.RequestAuth, stdReq promptcompat.StandardRequest) (promptcompat.StandardRequest, error) {
 	stdReq = shared.ApplyThinkingInjection(h.Store, stdReq)
 	svc := history.Service{Store: h.Store, DS: h.DS}
 	out, err := svc.ApplyCurrentInputFile(ctx, a, stdReq)
-	if err != nil {
-		return stdReq, err
+	if err != nil || out.CurrentInputFileApplied {
+		return out, err
 	}
-	if out.CurrentInputFileApplied {
-		return out, nil
-	}
-	return svc.Apply(ctx, a, out)
+	return out, nil
 }

 func (h *openAITestSurface) preprocessInlineFileInputs(ctx context.Context, a *auth.RequestAuth, req map[string]any) error {
@@ -114,8 +111,8 @@ func splitOpenAIHistoryMessages(messages []any, triggerAfterTurns int) ([]any, [
 	return history.SplitOpenAIHistoryMessages(messages, triggerAfterTurns)
 }

-func buildOpenAIHistoryTranscript(messages []any) string {
-	return promptcompat.BuildOpenAIHistoryTranscript(messages)
+func buildOpenAICurrentInputContextTranscript(messages []any) string {
+	return promptcompat.BuildOpenAICurrentInputContextTranscript(messages)
 }

 func writeOpenAIError(w http.ResponseWriter, status int, message string) {
--- a/webui/src/features/settings/HistorySplitSection.jsx
+++ b/webui/src/features/settings/HistorySplitSection.jsx
@@ -1,51 +1,9 @@
-export default function HistorySplitSection({ t, form, setForm }) {
+export default function CurrentInputFileSection({ t, form, setForm }) {
    return (
        <div className="bg-card border border-border rounded-xl p-5 space-y-4">
            <div className="space-y-1">
-                <h3 className="font-semibold">{t('settings.historySplitTitle')}</h3>
-                <p className="text-sm text-muted-foreground">{t('settings.historySplitDesc')}</p>
-            </div>
-            <div className="grid grid-cols-1 md:grid-cols-2 gap-4">
-                <label className="flex items-start gap-3 rounded-lg border border-border bg-background/60 p-4">
-                    <input
-                        type="checkbox"
-                        checked={Boolean(form.history_split?.enabled)}
-                        onChange={(e) => setForm((prev) => ({
-                            ...prev,
-                            history_split: {
-                                ...prev.history_split,
-                                enabled: e.target.checked,
-                            },
-                            current_input_file: {
-                                ...prev.current_input_file,
-                                enabled: e.target.checked ? false : Boolean(prev.current_input_file?.enabled),
-                            },
-                        }))}
-                        className="mt-1 h-4 w-4 rounded border-border"
-                    />
-                    <div className="space-y-1">
-                        <span className="text-sm font-medium block">{t('settings.historySplitEnabled')}</span>
-                        <span className="text-xs text-muted-foreground block">{t('settings.historySplitEnabledDesc')}</span>
-                    </div>
-                </label>
-                <label className="text-sm space-y-2">
-                    <span className="text-muted-foreground">{t('settings.historySplitTriggerAfterTurns')}</span>
-                    <input
-                        type="number"
-                        min={1}
-                        max={1000}
-                        value={form.history_split?.trigger_after_turns || 1}
-                        onChange={(e) => setForm((prev) => ({
-                            ...prev,
-                            history_split: {
-                                ...prev.history_split,
-                                trigger_after_turns: Number(e.target.value || 1),
-                            },
-                        }))}
-                        className="w-full bg-background border border-border rounded-lg px-3 py-2"
-                    />
-                    <p className="text-xs text-muted-foreground">{t('settings.historySplitTriggerHelp')}</p>
-                </label>
+                <h3 className="font-semibold">{t('settings.currentInputFileTitle')}</h3>
+                <p className="text-sm text-muted-foreground">{t('settings.currentInputFileDesc')}</p>
            </div>
            <div className="grid grid-cols-1 md:grid-cols-2 gap-4">
                <label className="flex items-start gap-3 rounded-lg border border-border bg-background/60 p-4">
@@ -54,10 +12,6 @@ export default function HistorySplitSection({ t, form, setForm }) {
                        checked={Boolean(form.current_input_file?.enabled)}
                        onChange={(e) => setForm((prev) => ({
                            ...prev,
-                            history_split: {
-                                ...prev.history_split,
-                                enabled: e.target.checked ? false : Boolean(prev.history_split?.enabled),
-                            },
                            current_input_file: {
                                ...prev.current_input_file,
                                enabled: e.target.checked,
@@ -89,7 +43,6 @@ export default function HistorySplitSection({ t, form, setForm }) {
                    <p className="text-xs text-muted-foreground">{t('settings.currentInputFileHelp')}</p>
                </label>
            </div>
-            <p className="text-xs text-muted-foreground">{t('settings.splitPassThroughHelp')}</p>
        </div>
    )
 }
--- a/webui/src/features/settings/SettingsContainer.jsx
+++ b/webui/src/features/settings/SettingsContainer.jsx
@@ -5,7 +5,7 @@ import { useSettingsForm } from './useSettingsForm'
 import SecuritySection from './SecuritySection'
 import RuntimeSection from './RuntimeSection'
 import BehaviorSection from './BehaviorSection'
-import HistorySplitSection from './HistorySplitSection'
+import CurrentInputFileSection from './HistorySplitSection'
 import CompatibilitySection from './CompatibilitySection'
 import AutoDeleteSection from './AutoDeleteSection'
 import ModelSection from './ModelSection'
@@ -96,7 +96,7 @@ export default function SettingsContainer({ onRefresh, onMessage, authFetch, onF

            <BehaviorSection t={t} form={form} setForm={setForm} />

-            <HistorySplitSection t={t} form={form} setForm={setForm} />
+            <CurrentInputFileSection t={t} form={form} setForm={setForm} />

            <CompatibilitySection t={t} form={form} setForm={setForm} />

--- a/webui/src/features/settings/useSettingsForm.js
+++ b/webui/src/features/settings/useSettingsForm.js
@@ -17,7 +17,6 @@ const DEFAULT_FORM = {
    responses: { store_ttl_seconds: 900 },
    embeddings: { provider: '' },
    auto_delete: { mode: 'none' },
-    history_split: { enabled: false, trigger_after_turns: 1 },
    current_input_file: { enabled: true, min_chars: 0 },
    thinking_injection: { enabled: true, prompt: '', default_prompt: '' },
    model_aliases_text: '{}',
@@ -52,8 +51,7 @@ function normalizeAutoDeleteMode(raw) {
 }

 function fromServerForm(data) {
-    const historySplitEnabled = Boolean(data.history_split?.enabled)
-    const currentInputFileEnabled = historySplitEnabled ? false : (data.current_input_file?.enabled ?? true)
+    const currentInputFileEnabled = data.current_input_file?.enabled ?? true
    return {
        admin: { jwt_expire_hours: Number(data.admin?.jwt_expire_hours || 24) },
        runtime: {
@@ -74,10 +72,6 @@ function fromServerForm(data) {
        auto_delete: {
            mode: normalizeAutoDeleteMode(data.auto_delete),
        },
-        history_split: {
-            enabled: historySplitEnabled,
-            trigger_after_turns: Number(data.history_split?.trigger_after_turns || 1),
-        },
        current_input_file: {
            enabled: currentInputFileEnabled,
            min_chars: Number(data.current_input_file?.min_chars ?? 0),
@@ -92,8 +86,7 @@ function fromServerForm(data) {
 }

 function toServerPayload(form) {
-    const historySplitEnabled = Boolean(form.history_split?.enabled)
-    const currentInputFileEnabled = historySplitEnabled ? false : Boolean(form.current_input_file?.enabled)
+    const currentInputFileEnabled = Boolean(form.current_input_file?.enabled)
    return {
        admin: { jwt_expire_hours: Number(form.admin.jwt_expire_hours) },
        runtime: {
@@ -108,10 +101,6 @@ function toServerPayload(form) {
        responses: { store_ttl_seconds: Number(form.responses.store_ttl_seconds) },
        embeddings: { provider: String(form.embeddings.provider || '').trim() },
        auto_delete: { mode: normalizeAutoDeleteMode(form.auto_delete) },
-        history_split: {
-            enabled: historySplitEnabled,
-            trigger_after_turns: Number(form.history_split?.trigger_after_turns || 1),
-        },
        current_input_file: {
            enabled: currentInputFileEnabled,
            min_chars: Number(form.current_input_file?.min_chars ?? 0),
--- a/webui/src/locales/en.json
+++ b/webui/src/locales/en.json
@@ -385,17 +385,11 @@
        "thinkingInjectionDesc": "Append a structured <think> checklist to the latest user message before prompt assembly.",
        "thinkingInjectionPrompt": "Thinking format prompt",
        "thinkingInjectionPromptHelp": "Leave empty to use the built-in default prompt shown as the input placeholder.",
-        "historySplitTitle": "Context Split",
-        "historySplitDesc": "Choose one context-splitting mode to avoid inlining very long prompts.",
-        "historySplitEnabled": "Turn split (second turn by default)",
-        "historySplitEnabledDesc": "After the configured user-turn threshold, pack earlier conversation into HISTORY.txt.",
-        "historySplitTriggerAfterTurns": "Trigger threshold (user turns)",
-        "historySplitTriggerHelp": "Default is 1, which means history split starts from the second turn.",
+        "currentInputFileTitle": "Independent Split",
        "currentInputFileEnabled": "Independent split (by size)",
-        "currentInputFileDesc": "After the character threshold is reached, upload the full context as a hidden context file and skip HISTORY.txt.",
+        "currentInputFileDesc": "Enabled by default. Once the character threshold is reached, upload the full context as a hidden context file.",
        "currentInputFileMinChars": "Current input threshold (characters)",
-        "currentInputFileHelp": "Default is 0, which uses independent split whenever there is input.",
-        "splitPassThroughHelp": "Turn split and independent split are mutually exclusive; choose at most one. If both are unchecked, requests pass through directly without uploading split context files.",
+        "currentInputFileHelp": "Default is 0, which uses independent split for any non-empty input.",
        "compatibilityTitle": "Compatibility",
        "compatibilityDesc": "Compatibility controls that keep stream output closer to the wire format or safer for the web UI.",
        "stripReferenceMarkers": "Strip [reference:N] markers",
--- a/webui/src/locales/zh.json
+++ b/webui/src/locales/zh.json
@@ -385,17 +385,11 @@
        "thinkingInjectionDesc": "在组装 prompt 前，将结构化 <think> 检查清单追加到最新用户消息末尾。",
        "thinkingInjectionPrompt": "思考格式提示词",
        "thinkingInjectionPromptHelp": "留空时使用内置默认提示词；默认内容会显示在输入框占位文本中。",
-        "historySplitTitle": "上下文拆分",
-        "historySplitDesc": "选择一种上下文拆分方式，减少超长 prompt 直接内联。",
-        "historySplitEnabled": "轮次拆分（默认第二轮）",
-        "historySplitEnabledDesc": "从配置的用户回合数之后，将更早的对话整理成 HISTORY.txt。",
-        "historySplitTriggerAfterTurns": "触发阈值（用户回合数）",
-        "historySplitTriggerHelp": "默认值为 1，表示从第二轮开始拆分历史。",
+        "currentInputFileTitle": "独立拆分",
        "currentInputFileEnabled": "独立拆分（按量）",
-        "currentInputFileDesc": "达到字符阈值后，将完整上下文上传为隐藏上下文文件，并跳过 HISTORY.txt。",
+        "currentInputFileDesc": "默认开启。达到字符阈值后，将完整上下文上传为隐藏上下文文件。",
        "currentInputFileMinChars": "当前输入阈值（字符数）",
-        "currentInputFileHelp": "默认 0，表示有输入时就使用独立拆分。",
-        "splitPassThroughHelp": "轮次拆分和独立拆分互斥，只能选择一种；如果都不勾选，请求会直接透传，不上传拆分上下文文件。",
+        "currentInputFileHelp": "默认 0，表示只要有输入就会使用独立拆分。",
        "compatibilityTitle": "兼容性设置",
        "compatibilityDesc": "用于控制输出格式兼容性，避免把模型原始流里的标记直接暴露到前端。",
        "stripReferenceMarkers": "移除 [reference:N] 标记",
@@ -1 +1 @@
 .1.0
 .1.1