MCP 服务器

wellread

Shared research cache for AI agents. Caches web research across sessions and users - hit means instant answer from verified sources, miss means your research saves the next dev's tokens. Semantic search with freshness tracking, gap detection, and real-time token measurement via JSONL. Free, open source.

README

wellread — Another dev already searched that.

The problem

❌ Your agent researches every technical question from scratch — 10-20 turns per query
❌ When it doesn't search, it hallucinates — outdated APIs, wrong examples, broken code
❌ Each turn re-sends your entire conversation history, and the cost compounds
❌ Thousands of devs burning tokens on the same questions, every day

The fix

Before your agent searches the web, wellread checks what other devs already found.

Hit → instant answer from verified sources. Zero web searches. One turn.
Partial → starts from what exists, only researches the gaps.
Miss → normal research, then saves it for the next person.

The compounding effect

	Without wellread	With wellread
Turn 1 (fresh session)	200K tokens · 10 turns · 67s	647 tokens · 1 turn · 28s
Turn 30 (~40K context)	1.2M tokens	647 tokens
Turn 100 (~150K context)	3.5M tokens	647 tokens
Turn 250 (~480K context)	11M tokens	647 tokens

The deeper your session, the more expensive research gets, and the more wellread saves.

Install

npx wellread

Restart your editor. That's it.

Update: npx wellread@latest · Uninstall: npx wellread uninstall

Singleplayer

Your own research comes back to you. No repeat searches, no hallucinations from stale training data — real sources, verified.

Multiplayer

27 devs already used that Auth.js research before you got here. One person researched, everyone benefits.

Freshness

Each entry knows how fast its topic changes:

Type	Fresh	Re-check	Re-research
Stable (React, PostgreSQL)	6 months	1 year	after
Evolving (Next.js, Bun)	30 days	90 days	after
Volatile (betas, pre-release)	7 days	30 days	after

When an agent re-verifies, the clock resets for everyone.

Privacy

Only generalized research summaries are shared. No code, no file paths, no credentials, no project names. Your agent strips everything private before saving.