Add Permafrost (cache-aligning proxy + Claude Code plugin) to Others#655
Open
jianzhichun wants to merge 1 commit into
Open
Add Permafrost (cache-aligning proxy + Claude Code plugin) to Others#655jianzhichun wants to merge 1 commit into
jianzhichun wants to merge 1 commit into
Conversation
A cache-aligning proxy + Claude Code plugin that freezes the prompt prefix so DeepSeek's automatic context cache always hits, via the Anthropic-compatible endpoint. Adds docs/permafrost (EN+CN) and a row in the Others table of both README.md and README_cn.md.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Adds Permafrost to the Others section (README.md + README_cn.md), with bilingual integration docs under
docs/permafrost/.Why it belongs here
Permafrost is built specifically around the DeepSeek API: it is a zero-dependency passthrough proxy (and Claude Code plugin) for DeepSeek's Anthropic-compatible endpoint (
https://api.deepseek.com/anthropic) that keeps the prompt prefix byte-stable so DeepSeek's automatic context caching keeps hitting — deterministic tool ordering, env-block freezing, nonce stabilization, canonical serialization, cold-anchor coalescing for parallel subagents, and an opt-in idle keepalive.Measured against the live DeepSeek API on real Claude Code traffic (reproducible via
e2e/run_claude_code.shin the repo): 66% cache hit rate / 64% lower cost on a multi-turn agentic task; coalescing takes concurrent cold fan-out from 0% → 73% hit. The repo also documents measured DeepSeek cache behavior (prefix-unit matching, header-fingerprint cache identity) indocs/e2e-findings.md.MIT licensed, no telemetry, runs entirely locally.