MiroThinker
MiroThinker is a deep research agent optimized for complex research and prediction tasks. Our latest models, MiroThinker-1.7, achieves 74.0 and 75.3 on the BrowseComp and BrowseComp Zh, respectively.
- 8.5k
- Python
- Apache-2.0
- 5 mo ago
MiroThinker is a deep research agent optimized for complex research and prediction tasks. Our latest models, MiroThinker-1.7, achieves 74.0 and 75.3 on the BrowseComp and BrowseComp Zh, respectively.
Convert Claude Code or Codex into the most intelligent Deep Research Agent. Collect, search, and synthesize web research into a persistent, searchable wiki that builds on itself. Hosted API + MCP: hyperresearch.ai
🏆 Top-1 on 5+ benchmarks | Web UI | Supports MiroThinker, Claude, Kimi, OpenAI
Give your Hermes agent the web as real sources, never a made-up answer — multi-provider search and extraction with an optional local, key-free DonSeTch option.
Multi-agent research execution system: paper reproduction, benchmark evaluation, AutoResearch and claim-evidence graphs
Open-source deep research for AI agents: 40 channels, 10+ Chinese sources.
A Codex/Claude skill for thermal-fluid mechanical engineering research, proposal development, technical writing, data analysis, presentations, and AI-assisted workflows.
OreoLook — an open-source, OpenAI-compatible agentic web search and deep-research engine with streaming, MCP tools, Redis memory, Qdrant retrieval, and PDF exports.
TreeThinkerAgent is a lightweight orchestration layer that turns any LLM into an autonomous multi-step reasoning agent. It supports multi-step planning, tool execution, and final synthesis while exposing the entire reasoning process as a tree you can explore.
Turn any topic into a complete research report using AI agents. ResearchForge autonomously plans tasks, searches the web, analyzes information, and generates actionable insights.
Give your agent the web as real sources, never made-up answers — an MCP server with 15 search and 9 extract providers plus an optional local, key-free DonSeTch integration.
A sourced web research agent on Cloudflare Workflows. Ask a question, follow evidence gaps, get a report with links.
A computer operated by agents. Turn Codex or Claude Code into a local workspace for research, reports, decks, web pages, images, and QA.
Owner-only, OpenAI-first hibernating Nous Hermes research agent with durable Sites state, Secure MCP Tunnel transport, disposable ChatGPT Work compute, native web research, and tool-less Codex OAuth inference.
Durable web-research agent built with Parallel and LiteLLM, running on Render Workflows with a FastAPI gateway
Agentic scientific research platform that autonomously plans, searches, analyzes, critiques, and writes cited research reports using multi-agent workflows, hybrid RAG, persistent memory, and knowledge graphs.
A local research agent that reads every source in a group you curate — RSS/Atom feeds, arXiv, GitHub releases, Hacker News, Reddit, RSSHub — and answers with a citation on every line plus a coverage report of what it couldn't reach. Built on a trimmed fork of the pi agent harness.
In this we implement research agent using agno with groq llm
Agentic research harness with a self-checking pipeline
Runs in the browser on Hugging Face
Successfully developed a LangGraph-powered autonomous research agent that searches multiple external sources in parallel, filters and deduplicates results, and exports structured reports as PDF and Markdown with session memory.
Agent skills for Claude Code and Codex: multi-source web research that outputs a readable report where every cited number, date and quote is verified against saved page copies. Source ladder, conflict resolution and reusable run folders for follow-ups.
Deep research for any AI agent, for about a penny. A receipt for every claim.
AI Agent that maintains a list of the biggest problems occupying the best minds.
Six specialist agents per mode, across four modes, research, debate and return a forensic verdict on why a company failed, in about 22 seconds on one AMD MI300X. AMD Developer Hackathon build.
Turns raw account signals into a cited buyer briefing and a first-contact plan.