Research agents
deep research and web research
assafelovic/gpt-researcher assafelovic/gpt-researcher is the clear leader by adoption and breadth of features for general-purpose deep-research agents; before choosing, pick based on a single constraint you care about most (self-hosting/privacy, extreme long‑context/verifiability, or tight domain integrations), because many projects specialize to solve those gaps.
Which one matches your setup?
Answer any of the questions — the shortlist updates as you go. Recommendations come from the capability passports below, nothing else.
Comparison matrix
Axes are extracted from each project's docs by our review pipeline; the maturity score is computed from stars, growth and commit activity — not an opinion. Click a column to sort.
| Runs in | Models | Context | Cost to run | |||||
|---|---|---|---|---|---|---|---|---|
⭐ 29.3k +102/7d | Web appCoding-agent pluginCLI | BYOKOpenAIAnthropicGemini | Multi-source context | Aggregation & filtering | ●●●●● | Self-hostable | ●●●●● | Your API key; cloud model costs (≈$0.4 per deep research with o3-mini as documented) |
⭐ 2.9k +89/7d | Coding-agent plugin | BYOKOpenAIAnthropicGeminiOllama / localFixed provider | Whole-library access | Citation gating & confirmations | ●●●●● | Local model support | ●●●●● | Your API key or local models |
⭐ 2.1k +52/7d | Coding-agent pluginCLI | AnthropicFixed provider | Web-only | Human-in-the-loop | ●●●●● | Cloud APIs | ●●●●● | Depends on model provider |
⭐ 1.1k +73/7d | CLIWeb app | BYOKOllama / local | Whole-project access | Clarifying prompts & review | ●●●●● | Fully local optional | ●●●●● | Your API key (pay-per-use) or free local models |
⭐ 1.1k +73/7d | Coding-agent pluginCLI | Anthropic | Whole-repo analysis | Severity gating & dedup | ●●●●● | Cloud LLM via session | ●●●●● | Runs in your Claude session — model usage billed by the provider |
⭐ 239 +3/7d | CLIWeb appCI | BYOKOpenAIFixed provider | Whole-library analysis | Rerank + quotas | ●●●●● | Self-hostable | ●●●●● | Your API key for LLM/embeddings (optional local-only use) |
⭐ 8.4k +2/7d | Web app | Fixed provider | Diff only | Trace collection | ●●●●● | Fully local | ●●●●● | Free; self-host models or use HuggingFace |
⭐ 3.1k | CLIWeb app | BYOKOpenAIAnthropicGemini | Diff + related files | None mentioned | ●●●●● | Fully local possible | ●●●●● | OpenRouter API key to start, or self-hosted local GPU |
⭐ 892 +7/7d | CLIWeb app | Fixed provider | Session-limited | Protocol enforced | ●●●●● | Fully local | ●●●●● | Free, self-host |
⭐ 423 +5/7d | Coding-agent pluginCLI | OpenAIAnthropicGemini | Whole-repo analysis | Gated workflow & verification | ●●●●● | Cloud API required | ●●●●● | Requires access to Codex or Claude (your provider access); local LaTeX required |
⭐ 76 +1/7d | CLIWeb appCIGitHub Action | BYOK | File-level only | Governance gating | ●●●●● | Cloud via API key | ●●●●● | Self-hosted; your API key for AI services |
⭐ 58 +1/7d | CLIWeb appCICoding-agent plugin | BYOKOpenAIFixed provider | Multi-module pipeline | Strict evidence gating | ●●●●● | Self-hostable | ●●●●● | Requires your LLM + search API keys |
Capability profiles
Six axes, 0–5 each. The shape tells you the strategy: a wide hexagon is a generalist, a spike is a specialist. Showing the 8 most established — the rest are in the full catalog.
An open-source multi-agent deep-research pipeline that combines web scraping, local document analysis, and MCP integrations to produce long, cited research reports.
Deep Zotero-native integration that provides grounded, citation-linked paper chat and library-wide agent actions (reads, writes, tagging, notes) not offered by generic agents.
Provides a structured two-phase (outline then deep investigation) human-in-the-loop research workflow packaged specifically as skills for Claude Code / OpenCode / Codex, including parallel web research modules and report generation.
A local-first, bring-your-own-keys AI research assistant that runs on your computer and bundles domain-specific scientific skills, workflows, and a living lab notebook.
Bundles a bounded, audit-ready review→decide→patch→recheck loop with ledgered issues and edit-safety safeguards, exposed as a Claude Code skill.
A local-first research agent that reads original source windows and runs an investigation loop to produce cited, source-grounded reports instead of one-shot RAG.
Provides open-source deep research agents with extremely long (256K) context windows and very high tool-call budgets, enabling long‑horizon, verifiable multi-step research workflows.
An open-source research-agent framework that reproduces state-of-the-art results on multiple agentic benchmarks while remaining fully runnable locally.
All repositories (13)
An autonomous agent that conducts deep research on any data using any LLM providers
MiroThinker is a deep research agent optimized for complex research and prediction tasks. Our latest models, M…
🏆 Top-1 on 5+ benchmarks | Web UI | Supports MiroThinker, Claude, Kimi, OpenAI
An open-sourced research agent system deeply rooted in your Zotero library.
Structured deep research skill for Claude Code/Open Code/Codex with human-in-the-loop control
An AI co-scientist running on your desktop. Claude Science but better.
Pre-submission AI review stress-test for research papers. A Claude Code skill: review, verdict, revise, verify…
🔬🦞 A self-evolving AI research colleague for scientists. 285 skills, zero hallucination, persistent memory.
A highly customizable agentic harness for arXiv-ready ML/AI review papers (and beyond). It drives agentic AI l…
A library-science-inspired personal knowledge management system with LLM agents
Literature-grounded research idea exploration for CLI agents. 文献驱动的研究选题与方向探索工具。
AI-native macro investment research infrastructure with native MCP, terminal CLI, agent runtime, and disciplin…
专利侵权分析系统 —— 输入专利公开号，产出竞品侵权分析报告；同时打包成 skill，可被任意 agent（dsh, codex, claudecode 等） 调用。
Hosted alternatives
If running your own reviewer is more ops than you want, these managed services cover the same job.
A hosted agent that browses, reads and writes a cited report from one prompt, so you get long-horizon research without running a retrieval-and-agent stack.
Try OpenAI Deep Research →Managed answer engine with a research mode that gathers and cites sources for you, the hosted alternative to standing up your own web-research agent.
Try Perplexity →A hosted research assistant focused on the academic literature: it finds papers, extracts findings and summarizes across them without a pipeline you maintain.
Try Elicit →