跳转至

API Changelog (中文 + English)

导航 / Navigation: 返回项目首页 | 文档首页 | 测试手册

Versioning Policy

中文： 当前项目处于快速迭代阶段，统一使用 v1 路由前缀，采用“文档先行 + 小版本增量兼容”的方式推进。

English: The project is in active iteration. We keep /api/v1/* as the stable prefix and evolve with additive, backward-compatible updates whenever possible.

2026-06-03 (Current)

Added: Auth APIs

  • POST /api/v1/auth/register
  • POST /api/v1/auth/login
  • GET /api/v1/auth/me

说明: - 登录成功返回 access_token (Bearer) - x-user-id 仍可作为兼容回退方式（建议迁移到 Bearer）

Added: Project Detail API

  • GET /api/v1/projects/{project_id}

说明: - 返回项目基础信息 + members - 需通过项目成员权限检查

Added: Project Member Management

  • POST /api/v1/projects/{project_id}/members
  • DELETE /api/v1/projects/{project_id}/members/{member_id}

Added: Chat Session History APIs

  • GET /api/v1/chat/projects/{project_id}/sessions
  • GET /api/v1/chat/sessions/{session_id}/messages

Added: Runtime Run Timeline APIs

  • GET /api/v1/chat/sessions/{session_id}/runs
  • GET /api/v1/chat/runs/{run_id}/events

Changed: Chat Send Message Response

  • Endpoint: POST /api/v1/chat/sessions/{session_id}/messages
  • New response field: run_id

示例响应:

{
  "session_id": "a-session-id",
  "role": "assistant",
  "text": "...",
  "run_id": "a-run-id"
}

Added: Worker Queue Mode for Memory Retry

  • Endpoint: POST /api/v1/memory/retry-embeddings
  • New query parameter: enqueue=true|false
  • Response新增字段：
  • queued
  • task_id
  • message

示例:

POST /api/v1/memory/retry-embeddings?limit=20&enqueue=true
Authorization: Bearer <token>

Backward Compatibility Notes

  1. x-user-id header fallback is still accepted by dependency resolver.
  2. Existing project/resource/chat core APIs remain available.
  3. New fields are additive; previous clients can ignore unknown fields safely.

Migration Checklist for Consumers

  1. 从 x-user-id 迁移到 Bearer token。
  2. 对 chat message 响应兼容 run_id 字段。
  3. 若接入任务队列，处理 memory retry 的 queued/task_id 返回值。
  4. 若前端需要可观测性，接入 runs/events API。

Next Planned Changes

  1. Runtime run 增加 token usage 与 latency 分段。
  2. Runtime events 增加 tool/mcp/memory 级别事件。
  3. Worker 增加任务状态查询 API（task list/detail/retry）。