@hifullmoon/aicommit 2.2.3 → 2.4.0

package/docs/provider-compatibility.md

@@ -1,25 +1,53 @@
 # Provider compatibility / Provider 兼容表
 
-Built-in setup defaults choose an adapter and endpoint; adapters own request/response dialects; the core owns Git state, user interaction, HTTPS enforcement, retry, timeout, authorization, and machine output.
+AICommit uses **Pi AI** (`@earendil-works/pi-ai`, pinned to 0.85.0) for model request construction, message conversion, SSE decoding, thinking events, and normalized results. Node.js **>=22.19.0** is required. See the [Pi AI documentation](https://github.com/earendil-works/pi/tree/main/packages/ai).
 
-内置 setup 默认值负责选择 adapter 和端点；adapter 负责请求/响应方言；核心负责 Git 状态、用户交互、HTTPS、重试、超时、鉴权与机器输出。
+AICommit 使用 **Pi AI**（`@earendil-works/pi-ai`，固定为 0.85.0）构造模型请求、转换消息、解析 SSE，并读取统一的推理事件和结果。要求 Node.js **>=22.19.0**。
 
-| Provider / adapter           | Endpoint and auth / 端点与鉴权                                              | Streaming / 流式                                                                | Reasoning / 推理                                                                        | Token and usage mapping / token 与 usage                                           | Notes / 说明                                                                                                 |
-| ---------------------------- | --------------------------------------------------------------------------- | ------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ |
-| OpenAI / `openai`            | Official HTTPS Chat Completions; Bearer key / 官方 HTTPS；Bearer key        | SSE                                                                             | Native for recognized `o*`/`gpt-5*`; model-dependent otherwise / 已识别推理模型原生支持 | `max_completion_tokens` for reasoning models, otherwise `max_tokens`; OpenAI usage | Unsupported effort is rejected locally for known model generations / 已知模型不支持的 effort 会本地拒绝      |
-| OpenRouter / `openrouter`    | OpenRouter HTTPS; Bearer key; `X-Title: aicommit`                           | SSE                                                                             | `reasoning.effort`                                                                      | `max_tokens`; OpenAI-style usage                                                   | Model IDs commonly include vendor prefix / model 通常含厂商前缀                                              |
-| DeepSeek / `deepseek`        | DeepSeek HTTPS; Bearer key                                                  | SSE                                                                             | `thinking.type` plus normalized effort / thinking 与归一 effort                         | `max_tokens`; compatible usage                                                     | `medium`/`xhigh` map to supported high behavior where required / 必要时映射为 high                           |
-| MiniMax / `minimax`          | MiniMax HTTPS; Bearer key                                                   | SSE                                                                             | `reasoning_split` and thinking switch                                                   | `max_tokens`; compatible usage                                                     | Adapter removes conflicting switches before send / 发送前移除冲突开关                                        |
-| Kimi Code / `custom`         | Kimi Code OpenAI-compatible HTTPS; Bearer key / Kimi Code OpenAI 兼容 HTTPS | SSE                                                                             | Server default; no vendor fields injected / 使用服务端默认值，不注入厂商字段            | `max_tokens`; OpenAI-style usage                                                   | Bundled preset uses `kimi-for-coding`; membership and Platform keys are distinct / 会员与开放平台 key 不通用 |
-| Ollama native / `ollama`     | Loopback `/api/chat` or `/api/generate`; normally keyless HTTP              | Complete JSON in v1; native NDJSON streaming is not consumed / v1 使用完整 JSON | Native `think` boolean                                                                  | `options.num_predict`; `prompt_eval_count` + `eval_count`                          | OpenAI-compatible `/v1/chat/completions` uses the compatible shape instead / `/v1` 使用兼容方言              |
-| Custom compatible / `custom` | HTTPS remote or loopback HTTP; optional core Bearer key                     | SSE when endpoint supports Chat Completions events                              | No vendor fields by default; explicit `enabledBody`/`disabledBody` / 默认不注入厂商字段 | `max_tokens`; common OpenAI/Anthropic/Ollama usage fields normalized               | Validate the endpoint before trusting it with code or credentials / 发送代码前验证端点                       |
+The runtime keeps the existing Provider/Model config format. `providers.js` selects Pi model metadata and applies configuration overrides; `model-client.js` calls Pi's `openai-completions` implementation. `provider-response.js` normalizes legacy response fields before Pi; its SSE framing uses the lightweight [eventsource-parser](https://github.com/rexxars/eventsource-parser) dependency. `api.js` retains commit prompts, policy validation, and recovery. Presets remain setup data, not executable adapters.
 
-## Compatibility contract / 兼容契约
+现有 Provider/Model 配置格式保持不变。`providers.js` 选择 Pi 模型元数据并应用配置覆盖；`model-client.js` 调用 Pi 的 `openai-completions` 实现；`provider-response.js` 在进入 Pi 前统一旧版响应字段，SSE 分帧使用轻量依赖 `eventsource-parser`；`api.js` 保留提交提示词、规则校验与恢复。预设仍是 setup 数据，不是可执行适配器。
 
-All built-in adapters return `content`, optional `reasoning`, normalized usage, finish reason, raw response, capabilities, attempts, and latency. Retries cover 429, selected 5xx, and network/body interruption; authentication, invalid parameters, and safety failures are not retried.
+| Provider / adapter        | Pi integration / Pi 接入                                                                                                                                 | Reasoning / 推理                                                                                                    | Token budget / 输出预算                                    |
+| ------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------- |
+| OpenAI / `openai`         | Chat Completions; bundled model metadata / Chat Completions 与内置模型元数据                                                                             | Pi thinking-level map; unsupported known efforts rejected locally / 使用 Pi 强度映射，已知不支持的强度本地拒绝      | Reasoning: `max_completion_tokens`; otherwise `max_tokens` |
+| OpenRouter / `openrouter` | Chat Completions; `X-Title: aicommit`                                                                                                                    | Pi `reasoning.effort` and catalog capability checks / Pi 参数映射与模型能力检查                                     | `max_tokens`                                               |
+| DeepSeek / `deepseek`     | Chat Completions; bundled model metadata / 内置模型元数据                                                                                                | Pi `thinking.type` and effort mapping; enabled thinking omits temperature / Pi 映射推理参数，开启时省略 temperature | `max_tokens`                                               |
+| MiniMax / `minimax`       | Chat Completions / 兼容接口                                                                                                                              | Small compatibility override for `reasoning_split` and thinking switches / 保留少量开关兼容逻辑                     | `max_tokens`                                               |
+| Kimi Code / `custom`      | Existing OpenAI-compatible endpoint and `kimi-for-coding` preset / 保留现有兼容端点与预设                                                                | Server defaults; configurable body switches / 服务端默认值或自定义请求体                                            | `max_tokens`                                               |
+| Ollama / `ollama`         | Native JSON bridge for `/api/chat` and `/api/generate`; compatible `/v1/chat/completions` uses Pi directly / 原生端点使用 JSON 桥接，兼容端点直接使用 Pi | Native `think` switch / 原生开关                                                                                    | Native: `options.num_predict`; compatible: `max_tokens`    |
+| Custom / `custom`         | Arbitrary OpenAI-compatible model IDs and full endpoint URLs / 任意兼容模型 ID 与完整端点 URL                                                            | `enabledBody` / `disabledBody`; no inferred vendor fields / 不自动注入厂商开关                                      | `max_tokens`                                               |
 
-所有内置 adapter 返回 `content`、可选 `reasoning`、标准 usage、finish reason、raw response、capability、attempts 与 latency。仅 429、部分 5xx、网络/响应中断会重试；鉴权、参数与安全错误不会重试。
+## Configuration and model metadata / 配置与模型元数据
 
-Use `aicommit doctor -p provider-name` to verify the selected adapter and live endpoint. Reuse a built-in adapter when only setup defaults change, and use `custom` for an OpenAI-compatible endpoint with optional body switches. A protocol requiring a different transport, streaming parser, or credential scheme needs core support.
+- `apiUrl` is still the **complete endpoint**, not a base URL. Proxy paths and query parameters are preserved. Requests use only the resolved AICommit credential. Pi environment-key discovery and OAuth are not invoked, and redirects are rejected.
+- `modelId` need not exist in Pi's catalog. Known OpenAI, DeepSeek, and OpenRouter models use the bundled metadata; unknown IDs retain a compatible fallback. No online model discovery runs during setup or generation.
+- `reasoning.mode: auto` preserves server defaults and explicit `extraBody`. `on` / `off` applies the selected mode after extras. Setup filters known supported effort levels. DeepSeek's legacy unsupported effort values are normalized by Pi: for the pinned V4 Flash catalog, `medium` becomes `high`, and `xhigh` becomes `max`.
+- Pi requests SSE by default, even when the CLI does not display reasoning. Complete JSON responses are bridged into Pi events. Set `extraBody: { "stream": false }` for endpoints that reject streaming requests; streaming-only options are removed automatically.
+- Native Ollama remains non-streaming. `/api/generate` receives `system` and `prompt`, while `/api/chat` receives `messages`; existing `options` are retained.
 
-使用 `aicommit doctor -p provider-name` 校验所选 adapter 与在线端点。若只改变 setup 默认值，应复用内置 adapter；OpenAI-compatible endpoint 及少量 body 开关使用 `custom`。需要不同传输、流解析或鉴权方案的协议必须由核心直接支持。
+对应行为：
+
+- `apiUrl` 仍填写**完整接口地址**，代理路径与查询参数会保留。鉴权只使用 AICommit 已解析的凭据，不调用 Pi 的环境变量凭据发现或 OAuth，也不跟随重定向。
+- Pi 目录中没有的 `modelId` 也可以配置；已知 OpenAI、DeepSeek、OpenRouter 模型使用随依赖发布的元数据，未知模型走兼容路径。setup 与生成过程不在线拉取模型目录。
+- `reasoning.mode: auto` 保留服务端默认值及显式 `extraBody`；`on` / `off` 在 extras 之后应用。setup 根据已知能力过滤强度。DeepSeek 的旧配置由 Pi 归一：当前 V4 Flash 目录中 `medium` 映射为 `high`，`xhigh` 映射为 `max`。
+- Pi 默认请求 SSE，包括终端不展示推理的场景；完整 JSON 响应通过桥接交给 Pi。若服务拒绝流式请求，可配置 `extraBody: { "stream": false }`，流式专用参数会自动移除。
+- Ollama 原生端点继续使用非流式响应，保留 `options`；`/api/generate` 使用 `system` / `prompt`，`/api/chat` 使用 `messages`。
+
+## Result and retry contract / 结果与重试契约
+
+Callers receive `content`, optional `reasoning`, normalized usage, finish reason, capabilities, attempts, and latency. Cached input tokens are included once in `inputTokens`; reasoning tokens are already part of output usage. `piMessage` exposes Pi's normalized assistant result. `raw` retains the actual complete JSON response, or a reconstructed Chat Completions result for SSE, preserving `callAPI` compatibility.
+
+Retries remain owned by AICommit; Pi and the underlying SDK's automatic retries are disabled. Only 429, selected 5xx, and network failures **before an accepted response** can retry. Accepted-body interruptions, malformed responses, authentication, invalid parameters, and safety failures are never automatically replayed. Oversized `Retry-After` fails instead of retrying early.
+
+SSE must contain a `finish_reason`. A clean EOF or `[DONE]` alone is rejected, and partial content is not returned as a successful generation. Token-limit aliases (`max_tokens`, `max_output_tokens`, `token_limit`) are normalized to `length` before Pi so recovery can run. Textual `reasoning_details`, including legacy shapes, are combined with ordinary reasoning deltas in arrival order. Duplicate representations within one event are emitted once; repeated text in later events is retained. Encrypted metadata remains opaque.
+
+业务仍获得 `content`、可选 `reasoning`、归一 usage、结束原因、能力、尝试次数与耗时。缓存输入 token 只计入一次，推理 token 已包含在输出中。`piMessage` 提供 Pi 的统一 assistant 结果；完整 JSON 的 `raw` 保留原响应，SSE 的 `raw` 则重建兼容 Chat Completions 结构。
+
+重试由 AICommit 负责，Pi 与底层 SDK 的自动重试均已关闭。只有 429、部分 5xx，以及**收到成功响应前**的网络失败可重试；已接受请求后的响应中断、格式错误、鉴权、参数及安全错误不会自动重放。超过上限的 `Retry-After` 会直接报错，不提前重试。
+
+SSE 必须带 `finish_reason`。只有 EOF 或 `[DONE]` 时会拒绝结果，不将半截内容当作成功生成。输出上限别名（`max_tokens`、`max_output_tokens`、`token_limit`）在进入 Pi 前归一为 `length`，保留补全恢复流程。`reasoning_details`（含旧格式）中的文本与普通推理分片按到达顺序合并；同一事件中的重复表示只展示一次，后续事件中重复出现的文本则保留。加密元数据不会作为推理文本输出。
+
+Use `aicommit doctor -p provider-name -m model-name` to verify a configured connection. This integration covers the existing six adapter types; installing Pi does **not** automatically expose every Pi provider, OAuth flow, or native Anthropic/Gemini/Responses endpoint. Those protocols require explicit routing and configuration support.
+
+使用 `aicommit doctor -p provider-name -m model-name` 检查配置的连接。本次接入覆盖现有六种适配类型；安装 Pi **不会自动开放**其全部供应商、OAuth 或 Anthropic/Gemini/Responses 原生端点，这些协议需要显式扩展路由与配置。

package/docs/troubleshooting.md

@@ -26,8 +26,20 @@ JSON 模式保证 stdout 只有一个机器对象，诊断进入 stderr。`error
 | `split run --scope=all --yes` stops before API call / split 非交互在 API 前停止 | `sensitive_data` / `7`          | Complete untracked scan found sensitive-looking data / 完整未跟踪扫描发现疑似敏感数据                                           | Review/stage intended files explicitly; do not bypass without checking the actual content                                                               |
 | Commit aborts after generation / 生成后提交中止                                 | `concurrent_modification` / `8` | Index/worktree changed during the protected window / 受保护窗口中 index/worktree 被修改                                         | Review `git status`, restore the intended snapshot, and generate again                                                                                  |
 | Split stopped after one or more commits / split 部分提交后停止                  | reported Git failure            | Hook, crash, SIGINT, or concurrent pending edit / hook、崩溃、中断或待处理文件变化                                              | Run `aicommit split resume`; if another Git workflow already replaced the transaction, use `aicommit split abort`（只删除恢复元数据，不改提交或工作区） |
+| `aicommit update` refuses the installation / update 拒绝当前安装                | `config` / `2`                  | Source checkout, npm link, or another active Node/npm environment / 源码、npm link 或当前是另一套 Node/npm 环境                 | Activate the Node environment that installed AICommit, or run `npm install --global @hifullmoon/aicommit@latest` manually                               |
+| `aicommit update` cannot reach npm / update 无法访问 npm                        | `network` / `4`                 | Registry authentication, proxy, DNS, or network failure / registry 鉴权、代理、DNS 或网络失败                                   | Check `npm config get registry` and npm authentication/proxy settings, then retry                                                                       |
 | npm provenance is absent or invalid / npm provenance 缺失或失败                 | npm audit failure               | Old npm CLI, non-trusted release, or wrong version / npm 过旧、非可信发布或版本错误                                             | Upgrade npm; run `npm audit signatures`; install only a version linked to the official workflow                                                         |
 
 If a failure remains, capture `aicommit doctor --output=json`, Node/Git versions, the error category, and redacted config sources. Never attach a diff, commit message, config file, API key, reasoning trace, or credential-helper output to a public issue.
 
 若问题仍未解决，请记录 `aicommit doctor --output=json`、Node/Git 版本、错误分类与脱敏后的配置来源。不要在公开 issue 中附加 diff、commit message、配置文件、API key、reasoning 或 credential-helper 输出。
+
+## Large-change limits / 大变更限制
+
+Default `auto` analysis builds a local inventory and selects bounded excerpts without model reduction calls. If a complete split candidate inventory cannot fit one request, stage a smaller logical change or explicitly choose personal `largeChange.strategy: "deep"`. Deep analysis can reach its fixed request (256) or depth (8) limits; increasing token or time budgets does not raise those limits. Token/time limits remain configurable in personal settings. Unknown model token counts use conservative estimates; an oversized request fails before dispatch, and provider-context errors are not blindly replayed.
+
+默认 `auto` 在本地建立清单并选择受限片段，不调用模型递归汇总。完整拆分候选清单放不进一次请求时，请暂存更小的逻辑变更，或明确在个人配置中选择 `largeChange.strategy: "deep"`。深度分析的请求数（256）和汇总层级（8）上限固定，提高 token 或时间预算不会改变它们；token 和时间预算可在个人配置中调整。未知模型采用保守 token 估算，输入估算超限会在发送前失败；Provider 上下文超限不会盲目重试。
+
+Text lines over 1 MiB fail explicitly. Large-change hunk plans are unsupported; use file-level split. Metadata-only files remain in the complete plan. Invalid, duplicate, or missing IDs are response-format errors, not automatic catch-all groups.
+
+单行超过 1 MiB 会明确报错。大变更暂不支持 hunk 规划，请使用文件级拆分。仅统计的文件仍纳入完整计划。无效、重复或遗漏 ID 会报响应格式错误，不会自动归入兜底组。

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@hifullmoon/aicommit",
-  "version": "2.2.3",
+  "version": "2.4.0",
   "description": "Safe, local-first AI commit message generator for Git workflows",
   "type": "module",
   "bin": {
@@ -48,6 +48,7 @@
     "lines": 70
   },
   "dependencies": {
+    "@earendil-works/pi-ai": "0.85.0",
     "@inquirer/checkbox": "^4.3.2",
     "@inquirer/confirm": "^5.1.0",
     "@inquirer/core": "^10.3.2",
@@ -57,6 +58,7 @@
     "@inquirer/select": "^4.1.0",
     "boxen": "^8.0.1",
     "chalk": "^5.4.0",
+    "eventsource-parser": "4.1.0",
     "ora": "^8.2.0",
     "wrap-ansi": "^9.0.2"
   },
@@ -90,7 +92,7 @@
     }
   },
   "engines": {
-    "node": ">=18"
+    "node": ">=22.19.0"
   },
   "devDependencies": {
     "c8": "10.1.3",

package/schemas/aicommit-output.schema.json

@@ -19,9 +19,15 @@
     "error"
   ],
   "properties": {
-    "schemaVersion": { "const": "1.0" },
-    "ok": { "type": "boolean" },
-    "message": { "type": ["string", "null"] },
+    "schemaVersion": {
+      "const": "1.0"
+    },
+    "ok": {
+      "type": "boolean"
+    },
+    "message": {
+      "type": ["string", "null"]
+    },
     "plan": {
       "type": ["array", "null"],
       "items": {
@@ -29,8 +35,15 @@
         "additionalProperties": false,
         "required": ["message", "files"],
         "properties": {
-          "message": { "type": "string" },
-          "files": { "type": "array", "items": { "type": "string" } },
+          "message": {
+            "type": "string"
+          },
+          "files": {
+            "type": "array",
+            "items": {
+              "type": "string"
+            }
+          },
           "hunks": {
             "type": "array",
             "items": {
@@ -38,29 +51,62 @@
               "additionalProperties": false,
               "required": ["path", "ids"],
               "properties": {
-                "path": { "type": "string" },
-                "ids": { "type": "array", "items": { "type": "string" } }
+                "path": {
+                  "type": "string"
+                },
+                "ids": {
+                  "type": "array",
+                  "items": {
+                    "type": "string"
+                  }
+                }
               }
             }
           }
         }
       }
     },
-    "provider": { "type": ["string", "null"] },
-    "model": { "type": ["string", "null"] },
-    "latencyMs": { "type": ["number", "null"], "minimum": 0 },
+    "provider": {
+      "type": ["string", "null"]
+    },
+    "model": {
+      "type": ["string", "null"]
+    },
+    "latencyMs": {
+      "type": ["number", "null"],
+      "minimum": 0
+    },
     "usage": {
       "type": ["object", "null"],
       "additionalProperties": false,
       "properties": {
-        "inputTokens": { "type": "number", "minimum": 0 },
-        "outputTokens": { "type": "number", "minimum": 0 },
-        "totalTokens": { "type": "number", "minimum": 0 }
+        "inputTokens": {
+          "type": "number",
+          "minimum": 0
+        },
+        "outputTokens": {
+          "type": "number",
+          "minimum": 0
+        },
+        "totalTokens": {
+          "type": "number",
+          "minimum": 0
+        }
       }
     },
-    "warnings": { "type": "array", "items": { "type": "string" } },
-    "exitReason": { "type": "string", "minLength": 1 },
-    "committed": { "type": "boolean" },
+    "warnings": {
+      "type": "array",
+      "items": {
+        "type": "string"
+      }
+    },
+    "exitReason": {
+      "type": "string",
+      "minLength": 1
+    },
+    "committed": {
+      "type": "boolean"
+    },
     "error": {
       "type": ["object", "null"],
       "additionalProperties": false,
@@ -78,9 +124,65 @@
             "internal"
           ]
         },
-        "message": { "type": "string" }
+        "message": {
+          "type": "string"
+        }
       }
     },
-    "data": { "type": "object" }
+    "data": {
+      "type": "object",
+      "properties": {
+        "analysis": {
+          "type": "object",
+          "description": "Large-change coverage and shared request budget; never contains diff or intermediate summaries.",
+          "properties": {
+            "totalFiles": {
+              "type": "number",
+              "minimum": 0
+            },
+            "analyzedFiles": {
+              "type": "number",
+              "minimum": 0
+            },
+            "sampledFiles": {
+              "type": "number",
+              "minimum": 0,
+              "description": "Files represented by partial excerpts; not fully analyzed files."
+            },
+            "strategy": {
+              "enum": ["auto", "deep"]
+            },
+            "metadataOnlyFiles": {
+              "type": "number",
+              "minimum": 0
+            },
+            "failedFiles": {
+              "type": "number",
+              "minimum": 0
+            },
+            "completedChunks": {
+              "type": "number",
+              "minimum": 0
+            },
+            "requests": {
+              "type": "number",
+              "minimum": 0
+            },
+            "budgetedTokens": {
+              "type": "number",
+              "minimum": 0
+            },
+            "maxTotalTokens": {
+              "type": "number",
+              "minimum": 0
+            },
+            "elapsedMs": {
+              "type": "number",
+              "minimum": 0
+            }
+          }
+        }
+      }
+    }
   }
 }

package/src/analysis-budget.js

@@ -0,0 +1,76 @@
+import { ERROR_CATEGORIES, fail } from './errors.js';
+
+export const DEFAULT_LARGE_CHANGE = Object.freeze({
+  strategy: 'auto',
+  chunkInputTokens: 12000,
+  maxTotalTokens: 200000,
+  concurrency: 2,
+  timeoutMs: 180000,
+});
+
+export function estimateTokens(text) {
+  // Deliberately conservative fallback for providers without a tokenizer.
+  return Math.ceil(Buffer.byteLength(String(text), 'utf8') / 2);
+}
+
+export function createAnalysisBudget(settings = {}) {
+  const limits = { ...DEFAULT_LARGE_CHANGE, ...settings };
+  const started = Date.now();
+  let charged = 0;
+  let requests = 0;
+  let inputTokens = 0;
+  let outputTokens = 0;
+  let reserveFinal = 0;
+  return {
+    limits,
+    signal: AbortSignal.timeout(limits.timeoutMs),
+    reserveFinal(tokens) {
+      reserveFinal = tokens;
+    },
+    remainingMs() {
+      return Math.max(0, limits.timeoutMs - (Date.now() - started));
+    },
+    snapshot() {
+      return {
+        requests,
+        budgetedTokens: charged,
+        maxTotalTokens: limits.maxTotalTokens,
+        elapsedMs: Date.now() - started,
+        usage: { inputTokens, outputTokens, totalTokens: inputTokens + outputTokens },
+      };
+    },
+    reserve(input, output) {
+      if (
+        !this.remainingMs() ||
+        charged + input + output + reserveFinal > limits.maxTotalTokens ||
+        requests >= 256
+      ) {
+        throw fail(
+          ERROR_CATEGORIES.PROVIDER,
+          'Large-change analysis reached its token, request, or time budget. No incomplete plan will be committed. Increase the personal largeChange budget or stage a smaller change.',
+          { data: { analysis: this.snapshot() } },
+        );
+      }
+      if (input > limits.chunkInputTokens) {
+        throw fail(
+          ERROR_CATEGORIES.PROVIDER,
+          'Analysis request exceeds largeChange.chunkInputTokens; shorten repository context or increase the personal input budget.',
+        );
+      }
+      charged += input + output;
+      requests += 1;
+      inputTokens += input;
+      outputTokens += output;
+      return { input, output };
+    },
+    settle(ticket, usage) {
+      if (!ticket || !usage) return;
+      const actualInput = usage.inputTokens ?? ticket.input;
+      const actualOutput = usage.outputTokens ?? ticket.output;
+      charged +=
+        Math.max(usage.totalTokens || 0, actualInput + actualOutput) - ticket.input - ticket.output;
+      inputTokens += actualInput - ticket.input;
+      outputTokens += actualOutput - ticket.output;
+    },
+  };
+}