#2323rd of 35 in Agent platforms
hive
Runtime where a queen agent spawns parallel worker agents sharing a SQLite tracker
- Stars
- 11.1k
- License
- Apache-2.0
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- Python
9,188 on 31 Mar 202611,113 on 20 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Hive runs a persistent "queen" agent that does one unit of a job itself, saves the method as a skill, then fans the rest out to parallel worker clones through tool calls rather than a drawn graph. Workers write results to a shared SQLite tracker the queen queries, and state is saved to disk so runs resume after a crash. It uses LiteLLM providers, MCP servers, and cron or webhook triggers.
Who it is for: Teams automating many similar business tasks with long-running, resumable agents
Strengths
- One agent-loop primitive; workers are clones, so no workflow graph to design
- Shared SQLite tracker lets runs be queried, audited and resumed from where they stopped
- Any LiteLLM provider, OpenAI-compatible endpoint, or local Ollama model; workers can use a different model
- Hard per-worker tool-call budgets, default four concurrent workers, and cron or webhook triggers
Weaknesses
- No dollar spending limits yet; only turn, tool-call and concurrency caps
- Repo is a uv workspace; pip install -e . does not work, quickstart script required
- Memory scores 88.4% on LongMemEval-S, below the 96.8% of reading full history
- No Docker or compose files detected; needs Python 3.11+ and Node.js 20+
What it needs
- GPU optional
- Needs Python 3.11+, Node.js 20+, git, uv, ripgrep, Ollama (optional, for local models)
- Models: Anthropic, OpenAI, Google Gemini, Groq, Cerebras
- port 8787
Also in Agent platforms
See all 35| Rank | Project | Score |
|---|---|---|
| 1 | LangflowVisual builder for AI agents and workflows, exposed as API or MCP server | 81 out of 100 |
| 2 | daily_stock_analysisDaily LLM stock analysis for A-share, HK, US and other markets, pushed to chat apps | 80 out of 100 |
| 3 | deer-flowAgent harness that runs sub-agents, memory and sandboxed skills for long tasks | 77 out of 100 |
| #4 | vibe-tradingFinance research agent for backtests, market data and broker connectors | 77 out of 100 |
| #5 | MulticaIssue board where AI coding agents take assignments like teammates | 76 out of 100 |