#8787th of 389 in Skills, plugins and prompts
billion-context
Context-compression proxy for coding agents running long sessions
- Stars
- 6.7k
- License
- custom license
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- TypeScript
350 on 4 Aug 20266,744 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
billion-context (the `bili` command) sits between a coding agent and its Anthropic or OpenAI-style model API and rewrites the stream with acp-kernel compression. It injects four tools (compress, decompress, search_context, acp_status) so the model folds old message ranges into summaries and restores them when needed. It installs as a native plugin, a launcher, or a `/bili/` base-URL prefix for clients such as pi, opencode, claude, codex, kimi and aider.
Who it is for: Heavy users of CLI coding agents who hit context limits and token costs
Strengths
- Compression is reversible: decompress restores exact details, and search_context covers summaries
- Works with about 25 clients via native plugin, launcher or base-URL prefix
- README reports 95-97% prefix-cache hit rate, with compression costing at most 2%
- Ships a paper and third-party benchmarks against full context and three RAG arms
Weaknesses
- Native plugin mode covers only some clients; others are launcher or cert-MITM only
- Running a standalone extension alongside a native plugin double-compresses context
- License detected as NOASSERTION, so terms need checking before use
- Setup is involved: ports, lanes and per-client modes require reading CLIENTS.md
What it needs
- no GPU
- Docker
- Needs Node.js, npm
- Models: Anthropic API, OpenAI API
- port 8787
Quick answers
Is billion-context open source?
Its code is public on GitHub under its own license terms; read them before you use it at work.
Can you run billion-context with Docker?
Yes. Its repository has a Docker setup: a Dockerfile, a Compose file or a published image.
Also in Skills, plugins and prompts
See all 389| Rank | Project | Score |
|---|---|---|
| 1 | claude-memPersistent cross-session memory plugin for Claude Code and other coding agents | 97 out of 100 |
| 2 | impeccableDesign skill with 24 slash commands for AI coding agents' frontend output | 97 out of 100 |
| 3 | gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles | 96 out of 100 |
| #4 | rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex | 96 out of 100 |
| #5 | cavemanShortens coding-agent replies and compresses tool output before the model reads it | 95 out of 100 |