Research and task-specific harnesses
Deep research, document QA, and domain-specific agent loops.
| # | Project | Stars | Tier | OSS | Description |
|---|---|---|---|---|---|
| 1 | DeerFlow memorymulti-agentsandboxpython | 84.8k | complex | ✅ | ByteDance's long-horizon research **harness**: a LangGraph-based agent loop that researches, codes, and creates using sandboxes, memory, tools, and subagents behind a message gateway—an alternative to the reference gpt-researcher stack for multi-hour tasks. |
| 2 | gpt-researcher multi-agentpython | 30.4k | complex | ✅ | Autonomous deep-research agent: web + local sources, citation-grounded reports, multi-agent and deep-research modes. The reference open-source research harness. |
| 3 | AutoResearchClaw multi-agent | 14.8k | complex | ❓ | Fully autonomous, self-evolving research **harness**: multi-agent debate and citation verification carry a project from idea to a written paper without a human in the loop. |
| 4 | MiroThinker evals | 8.4k | slightly complex | ❓ | Deep-research **harness** tuned for long browsing-and-reasoning chains; benchmarked on BrowseComp, GAIA, and HLE by pairing a dedicated agent loop with its own MiroThinker models rather than bolting search onto a generic chat agent. |
| 5 | openagents | 461 | complex | ✅ | Platform for autonomous agents and autopilot-style workflows; decentralized/Nostr-oriented (Pylon runtime, actively shipped in 2026). |
| 6 | Jev Social browserclitypescript | 194 | mostly simple | ✅ | Local-first social media research agent for Instagram, TikTok, and LinkedIn. The **harness** is a bounded decide-act-observe loop: the agent picks a typed, read-only operation from the currently valid candidates, the socai CLI runs it in the user's own Chrome session and returns evidence, and the run ends on an explicit finish or a step cap, keeping partial evidence and recoverable checkpoints when access is blocked. |