anyclaude-sdk clones the Claude Agent SDK interface for any API
A Show HN project called anyclaude-sdk rebuilds Claude Code’s agent harness on top of any OpenAI- or Anthropic-compatible endpoint. It keeps the same async query() generator and the same SDKMessage envelope as the official SDK, so code written against that SDK runs unchanged.
It runs in the browser via WebContainer, plus Node and Bun, with no backend, no OAuth, and no native binaries. The feature list covers MCP servers both external and in-process, sub-agents, agent teams with a coordinator, slash commands, and background tasks.
Three transport clients ship with it: OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages. Sandbox options are WebContainer, E2B, Vercel Sandbox, Daytona, Cloudflare Sandbox, and local. Install with npm install anyclaude-sdk, MIT licensed.
Up next
Letting Claude filter its own tool output took Opus 4.6 from 45.3% to 61.6% on BrowseComp
Anthropic: after 14,000 steps in Pokemon, Sonnet 3.5 had 31 sprawling note files and was stuck in the second city. Opus 4.6 at the same step had 10 files sorted into folders, three badges, and a file of lessons from its own failures.
Related stories
- Intent-lens spells out Claude Code calls before they run
- Codedict cuts Claude Code tokens 63%, but cost only 26%
- claude-files-pane shows Claude Code’s file reads live
- lcc cut one Claude Code session 49.5% without a summary
- catnap keeps Claude Code from compacting mid-task
- Script checks whether your network breaks Claude API uploads
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.