Oxynote
Oxynote is an open-source, self-hosted workspace for engineering documentation that puts live Prometheus and SQL metric charts alongside the pages explaining the systems. It combines collaborative pages, operational/API documentation, agent access, and pull-request-style draft, diff, and approval workflows in one workspace.
At a Glance
- Platform engineering and infrastructure teams
- Backend and service engineering teams
- Organizations running Prometheus or SQL-queryable operational data
- Teams seeking self-hosted alternatives to Notion, Confluence, Grafana, or Datadog for engineering context
- +1 more
AI Tools by Oxynote
(1)
Oxynote
Self Hosted Engineering Docs Wiki
Discussions
No discussions yet
Be the first to start a discussion about Oxynote
Latest News
Repository activity added Supabase as a data source and CA certificates to data sources.
PostgreSQL published “Oxynote: an open source internal wiki with live PostgreSQL query charts.”
Oxynote v0.9.3 released with split-documentation submenu improvements, reviewability safeguards, UI improvements, and authentication fixes.
Oxynote v0.9.2 released with tag-name and tag-color editing.
Products & Services
Apache-2.0 open-source technical wiki/workspace distributed as a Docker container or runnable from source. It provides collaborative engineering pages, live metric charts, API documentation, search, reviews, and optional AI/MCP capabilities.
An all-in-one production image containing the Caddy front door, Nuxt SSR web app, core API, auth/realtime service, and PostgreSQL. It runs on a customer’s own server; external PostgreSQL, Valkey/Redis, S3-compatible storage, SMTP, GitHub/Slack apps, and changedetection.io are optional.
The public GitHub repository contains the source, Apache-2.0 license, Docker quick start, production-image README, configuration documentation, and contribution guidance.
Market Position
Oxynote positions itself between an engineering wiki and observability tooling: unlike general-purpose Notion or Confluence, it natively embeds live Prometheus/SQL charts and reviews chart configuration with explanatory text; unlike Grafana or Datadog, it puts documentation, runbooks, API context, and pull-request-style content review beside the metrics. Its differentiators are self-hosting, Apache-2.0 licensing, one-container deployment, live metrics in documents, block-aware review, and MCP agent access. It remains early-stage: logs, traces, alerting, uptime monitoring, database viewing, and admin actions are described as in development.
Leadership
Founders
David (@davseby)
Co-founder of Oxynote. The project’s founding account describes him as having years of Grafana and internal-documentation experience; he previously worked on systems handling financial-market data over WebSockets and maintains the Go ttlcache project. His GitHub profile identifies him as David and links to LinkedIn (Dovydas Bykovas).
@swithek
Co-founder of Oxynote. The Oxynote README identifies @swithek as David’s co-founder and says the project comes from the authors of ttlcache; the GitHub profile for @swithek shows a Europe-based open-source developer and contributor to ttlcache-related work.
Executive Team
David (@davseby)
Co-founder and maintainer
GitHub identifies him as David; he has years of Grafana/internal-documentation experience and is associated with the ttlcache project.
@swithek
Co-founder and maintainer
Co-founder named in the Oxynote README and an open-source developer associated with ttlcache and the Oxynote repository.
Founding Story
The founders saw engineering teams split operational context across Notion or Confluence, Grafana or Datadog, terminals, and chat. David’s experience maintaining market-data services and later handling an incident where documentation and monitoring failed to connect reinforced the need for a single page containing the metric chart, explanation, command, and feature flag. Oxynote was started to make those elements editable together, reviewable like a pull request, and usable by both tired humans and AI agents; its original working idea was “breathing docs.”
Business Model
Revenue Model
The currently available product is free, Apache-2.0-licensed, self-hosted software with no per-seat fees. Oxynote runs on the customer’s own infrastructure; a hosted cloud version is planned but is not yet available.
Pricing Tiers
Apache-2.0 licensed software; deploy with Docker or from source on the customer’s own servers, with no member limit in a self-hosted install.
Target Markets
- Platform engineering and infrastructure teams
- Backend and service engineering teams
- Organizations running Prometheus or SQL-queryable operational data
- Teams seeking self-hosted alternatives to Notion, Confluence, Grafana, or Datadog for engineering context
- Organizations experimenting with MCP-compatible engineering agents
- Platform and infrastructure teams maintaining runbooks for services they operate
- Engineering teams keeping architecture notes, API references, on-call steps, and release checklists with live system context
- Internal API documentation that combines request/response fields, examples, and live endpoint metrics
- Incident response and operational troubleshooting
- Teams building MCP-based AI agent workflows for documentation and operations
- Solo engineers or small teams wanting a private self-hosted engineering wiki