Oynix
A knowledge-graph memory layer that gives AI coding agents persistent, factual context about your codebase, tickets, and docs — without another hosted service.
🔗 Visit Oynix
Description
AI coding assistants are great at writing a function, but the moment you ask 'who owns this file' or 'what would break if I change this,' most of them either guess or make something up, because they don't actually track how your code, tickets, and docs connect to each other over time. Oynix is built to fix that gap: it keeps a living map of your project — who touched what, which functions call which, what decisions were made and why — and hands that map to your AI agent as plain facts instead of a guess.
Under the hood, Oynix is a knowledge-graph server that ingests code, tickets, documentation, and communication threads and exposes them to AI agents over the MCP protocol. It runs on your own infrastructure with data in your own database (no hosted service required), tracks ownership and call/import edges from static source analysis, supports impact/blast-radius analysis before a change is made, and deliberately keeps models out of the retrieval path — it returns raw graph facts rather than an LLM's summary of them, which reduces hallucination risk for agents that consume it.
💬 Our review
The short version: a smart, narrow bet — instead of another RAG-over-your-code product, Oynix builds an actual structural graph of ownership and dependencies, which is the kind of grounded context that makes AI coding agents meaningfully more reliable, and the local-first, free-forever single-machine tier makes it easy to try with zero commitment.
Pricing scales purely with number of users/keys rather than repo size or commit volume ($15/mo solo, $99–$459/mo for teams), which is fair for larger codebases but means a growing team pays linearly regardless of how much value they're actually extracting. The 'no model on the retrieval path' design is a genuine differentiator versus typical vector-search RAG tools — you get deterministic graph facts, not a paraphrase — but that also means Oynix is infrastructure you plug into an agent, not a finished product with its own UI; you're trusting a young project to sit in your dev pipeline and keep the graph accurate as your code changes.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Local (1 machine) gratuit à vie sans compte. Solo 15$/mois (1 clé), Team 99$/mois (10 clés) ou 459$/mois (50 clés), Entreprise sur devis.
Pros
Graphe de connaissance réel (ownership, dépendances) plutôt qu'une simple recherche vectorielle
Auto-hébergé, données dans votre propre base — pas de service tiers obligatoire
Aucun LLM sur le chemin de restitution : faits bruts, pas de paraphrase
Palier local gratuit à vie pour une seule machine
Cons
Tarification par utilisateur/clé, pas par taille de dépôt — coûte plus cher à mesure que l'équipe grandit
Pas de produit fini avec interface — s'intègre à un agent existant via MCP
Projet jeune, à valider sur la fiabilité du graphe dans le temps