Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
7/24 Office — Self-evolving AI Agent system. 36 tools, 10,000 lines pure Python, modular architecture, MCP plugins, three-layer memory, nudge system, AI mirror, 24/7 production.
| Date | Stars |
|---|---|
| 2026-07-31 | 1031 |
| 2026-08-06 | 1031 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# 7/24 Office -- Self-Evolving AI Agent System
A production-running AI agent built in **~10,000 lines of pure Python** with **zero framework dependency**. No LangChain, no LlamaIndex, no CrewAI -- just the standard library + a few small packages.
**36 tools. 20 files. Modular architecture. Runs 24/7.**
Built solo with AI co-development tools. Production 24/7 across multiple users.
## Successor: xiaowang-v2 (Node rewrite, 2026)
The next generation lives in [`xiaowang-v2/`](./xiaowang-v2/): a single Node process + a single WAL SQLite + a hand-written agentic loop, zero framework dependency, running 24/7 over WeCom as a personal agent. Where 7/24 Office grew by adding tools, v2 converges: durable cross-day execution, layered memory with honest forgetting, turn assembly for chat-native input (burst messages / photo-then-caption / voice), and structural instruction-following — the model does fuzzy routing, the harness does the deterministic five (execution feedback, schema locks, state exposure, approval gates, idempotency). Design blueprints and 330+ offline selftest assertions included. Docs are in Chinese.
## What's New in v2.0
- **Modular tool architecture** -- Split from monolithic `tools.py` into 7 domain modules
- **Group chat support** -- Independent container for group conversations with @-mention gating
- **AI Mirror** -- Behavioral profile reports (`soul_report`) + future-self dialogue mode (`future_self`)
- **Nudge system** -- Structural behavior correction: auto-detects when LLM has tools but doesn't use them
- **Dynamic tool filtering** -- 5 context profiles (voice/scheduler/group/diagnostic/default) to reduce token waste
- **Budget-aware system prompt** -- Token budget tracking during system prompt assembly
- **Inactivity guard** -- Auto-skip cron tasks for dormant users (3-day threshold)
- **Circuit breaker** -- Disable tools after 3 consecutive failures per session
- **Interactive visualization** -- ECharts-based HTML pages via `render_page` (line/bar/pie/radar/table/report)
- **Container reconciliation** -- Router auto-rebuilds missing containers from routing table on startup
- **Exponential backoff retry** -- Messaging API calls retry 3x with 2/4/8s delays
- **Session auto-archiving** -- Daily black box recording of all conversations
## Features
- **Tool Use Loop** -- OpenAI-compatible function calling with automatic retry, up to 20 iterations per conversation
- **Three-Layer Memory** -- Session history + LLM-compressed long-term memory + LanceDB vector retrieval
- **MCP/Plugin System** -- Connect external MCP servers via JSON-RPC (stdio or HTTP), hot-reload without restart
- **Runtime Tool Creation** -- The agent can write, save, and load new Python tools at runtime (`create_tool`)
- **Self-Repair** -- Daily self-check, session health diagnostics, error log analysis, auto-notification on failure
- **Cron Scheduling** -- One-shot and recurring tasks, persistent across restarts, timezone-aware, inactivity guard
- **Multi-Tenant Router** -- Docker-based auto-provisioning, one container per user, health-checked, reconciliation
- **Multimodal** -- Image/video/file/voice/link handling, ASR (speech-to-text), vision via base64
- **Web Search** -- Multi-engine (Tavily, Bocha, GitHub, HuggingFace) with auto-routing and dual-engine default
- **Video Processing** -- Trim (with intelligent silence detection), add BGM, AI video generation via API
- **Messaging Integration** -- Pluggable messaging platform with debounce, message splitting, streaming media upload
- **Group Chat** -- Independent container, @-mention gating, context buffer (last 20 messages), speaker identification
## Architecture
```
+-----------------+
| Messaging |
| Platform |
+--------+--------+
|
+--------v--------+
| router.py | Multi-tenant routing
| Auto-prExcerpt of 11,493 characters
Read on GitHub
4
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:fdb958c281a5e9f2, desc:ai agent