Data / Repositories / Token optimization
Token optimization repositories on GitHub: 9 ranked by stars
9 public repositories tagged Token optimization are on GitHub, with 167,195 stars between them; 8 had a push in the last 30 days. Most are written in Python (5), Rust (2), TypeScript (1). The most starred: rtk-ai/rtk (82,881), headroomlabs-ai/headroom (75,347), yvgude/lean-ctx (3,878).
SpecStream is the live web data engine for teams and agents. It reads open-source repositories on GitHub and keeps what it reads, and this page groups that reading by topic. Ask for the repositories you track by topic, language, license or activity, and new ones and changes arrive in your sheet, CRM, Slack or agent.
Every repository, most starred first
| Repository | What it is | Language | Stars | Forks | License | Last push | Read |
|---|---|---|---|---|---|---|---|
| rtk-ai/rtk | CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies | Rust | 82,881 | 5,258 | Apache-2.0 | Oct 9 | Oct 11 |
| headroomlabs-ai/headroom | Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP… | Python | 75,347 | 5,847 | Apache-2.0 | Oct 13 | Oct 14 |
| yvgude/lean-ctx | LeanCTX — Context Gateway for AI Systems. Control what your AI can see. Open-source Engine for context selection, supported controls, and evidence. | Rust | 3,878 | 353 | Apache-2.0 | Oct 14 | Oct 14 |
| jgravelle/jcodemunch-mcp | Cut AI token costs 95%+ on code exploration. The leading MCP server for precise, symbol-level GitHub code retrieval via tree-sitter AST. Works with Claude Code, Cursor & any MCP… | Python | 2,769 | 375 | Oct 14 | Oct 14 | |
| Paritok-official/paritok-4b-v1 | Non-destructive compression gateway for AI coding agents. Cuts token bills 25% on turn 1 to past 85% in long or saturated sessions, and fits ~3× more turns in the same context… | Python | 1,459 | 140 | Apache-2.0 | Oct 6 | Oct 14 |
| elusznik/mcp-server-code-execution-mode | An MCP server that executes Python code in isolated rootless containers with optional MCP server proxying. Implementation of Anthropic's and Cloudflare's ideas for reducing MCP… | Python | 340 | 33 | GPL-3.0 | Dec 5, 2025 | Oct 14 |
| chipfighter/context-attachment-manager | Codex and Claude Code plugin (works in Claude Desktop): choose which past images go to the model with your next message. Unchecked images become placeholders, the conversation… | TypeScript | 235 | 22 | MIT | Oct 15 | Oct 15 |
| jgravelle/jdocmunch-mcp | The leading, most token-efficient MCP server for documentation exploration and retrieval via structured section indexing | Python | 204 | 48 | Oct 14 | Oct 14 | |
| mcpdelta/mcpdelta | Delta MCP is a free app that sits between your AI apps and their MCP servers: any server, every tool. It checks each task before it runs, applies it all or nothing, reads back… | 82 | 12 | Oct 14 | Oct 21 |
By language
- Python 5
Related topics
- Claude Code 8
- MCP 7
- Developer tools 6
- LLM 6
- AI coding 5
- Cursor 5
- Claude 5
- Model Context Protocol 5
- Anthropic 4
- Gemini CLI 4
- MCP server 4
- Agentic coding 3
Read from GitHub by SpecStream between October 11, 2026 and October 21, 2026; each row shows the day its repository was read. Stars and pushes change every day, and the GitHub page is the current one. Forks and archived repositories are left out.