About Us
Open source tools for grounding AI
Neuledge builds developer tools that connect AI agents to real, verified data — so they stop hallucinating and start being useful.
The Problem
AI coding assistants are powerful — until they hallucinate an API that doesn't exist, reference a deprecated method, or invent configuration options. The root cause is simple: LLMs don't have access to the data they need.
Cloud documentation services partially solve this, but they introduce their own problems — rate limits that cut off your agent mid-session, latency that slows down workflows, and recurring costs that scale with usage.
We're building the tools to fix this. Local-first, open source, and designed for developers who need reliable AI agents — not another SaaS subscription.
Why Open Source
AI grounding infrastructure should be transparent, auditable, and community-owned. Here's why we build in the open.
Trust Through Transparency
When your AI agent's accuracy depends on the tools feeding it data, you need to be able to inspect and verify those tools. Closed-source grounding is a black box you can't debug.
No Vendor Lock-In
Your documentation indexes and data layers run on your infrastructure. If Neuledge disappeared tomorrow, your tools keep working. That's the point.
Community-Driven
The best developer tools are shaped by the developers who use them. Open source means every user can file issues, submit fixes, and influence the roadmap.
What We Build
Two open source libraries that ground AI agents in different types of data.
@neuledge/context
Local-first documentation for AI agents. Indexes library docs into portable SQLite files and serves them via MCP. Sub-10ms queries, works offline, completely free.
Learn more →
@neuledge/graph
Semantic data layer for AI agents. Gives agents structured access to
live data sources through a single lookup() tool.
Pre-cached, under 100ms, structured JSON.
Learn more →
Built in the Open
All our tools are Apache 2.0 licensed and available on GitHub. Star, fork, or contribute — we'd love to hear from you.
View on GitHub