Neuledge Context
A free, offline library of documentation that AI coding assistants can search instantly, so they stop guessing about which version of a framework you're actually using.
🔗 Visit Neuledge Context
Description
AI coding assistants often give confidently wrong answers about a library because they were trained on some average of every version of its documentation ever published — so they'll suggest a method that was removed two major versions ago. Neuledge Context solves this by giving the assistant a real, version-matched copy of the documentation to search, stored locally rather than fetched live from the internet.
Neuledge Context is an open-source (Apache 2.0) tool that packages documentation for 100+ frameworks and libraries (React, Django, Spring Boot, and others) into portable SQLite databases with full-text search, indexed via SQLite FTS5 for sub-10ms query times. It integrates with AI assistants through MCP, works entirely offline with no cloud calls, no API keys, and no rate limits, and supports matching documentation to the exact version a project depends on. Teams can also index and search their own private documentation using the same mechanism.
💬 Our review
The short version: if your AI coding assistant keeps hallucinating outdated API methods, Neuledge Context is a free, zero-friction fix — plug it in via MCP and the assistant gets grounded, version-correct documentation instead of guessing from training data.
Compared to just letting an assistant search the web for docs (slow, and it can grab the wrong version) or relying on the assistant's built-in knowledge (often stale), Neuledge's local SQLite indexing is both faster and more accurate for version-sensitive questions — sub-10ms queries with no network round-trip. Being completely free and open source with no sign-up removes basically every reason not to try it, though the tradeoff is that with 100+ pre-built packages, less popular or very new libraries may not have a package ready yet, and private documentation still needs to be indexed manually. Solid, low-risk addition to any AI-assisted dev workflow.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Entièrement gratuit, sans limite d'usage, sans clé API, sans inscription. Licence Apache 2.0 permettant un usage commercial illimité.
Pros
Entièrement gratuit et open source (Apache 2.0)
Fonctionne hors-ligne, aucun appel cloud ni limite de débit
Requêtes en moins de 10ms grâce à l'indexation SQLite FTS5
Documentation adaptée à la version exacte des dépendances du projet
Cons
Bibliothèques moins connues ou très récentes pas forcément déjà packagées
Documentation privée à indexer manuellement
Projet encore jeune (créé février 2026), 358 étoiles GitHub
Dépend de l'intégration MCP côté assistant IA