Every repository I have starred, 1,509 of them. 54% is AI and agent tooling, which says more about the last two years than about the list. Search it, filter it, or paste your project into the matcher.
AI & Agents holds 808 of 1,509 repos, 54% of the index
my-starred-repos is a free, searchable index of 1509 GitHub repositories starred by Akash Priyadarshi: AI agents, MCP servers, Rust CLIs, RAG frameworks and developer tools, sorted into 11 categories. The page reads a baked snapshot of the public GitHub API, regenerated on demand, not a live feed. This build dates from 2026-11-17.
Every entry is here because it was useful for production work, architectural clarity, or agentic workloads. Nothing is sponsored and nothing is ranked by anyone but the star count, which is GitHub's.
my-starred-repos is a free index of every GitHub repository starred by Akash Priyadarshi, searchable by keyword, category, and language, with a client-side Project Matcher that ranks tools against your own project description. New stars appear on the next sync run.
Top-rated repositories include multi-agent orchestration frameworks (LangGraph, CrewAI, AutoGen, Hermes-Agent), Model Context Protocol servers (MCP), agent skill ecosystems (Claude Code skills, Superpowers, ECC), type-safe LLM tools (Pydantic-AI), token-efficient CLI utilities (rustygrep, zcat, repomap), and local agent assistants (OpenClaw, Browser-Use).
It scores every repository in the browser by keyword overlap across names, topics, languages, and descriptions, with a small boost for well-starred projects. Nothing you paste leaves the page, and scores are relative to the best match in the current dataset.
There is no schedule. The owner re-runs the sync script against the public GitHub API and commits the result; the snapshot label in the header shows how old the current build is.