# Hydron Roadmap > What Hydron is building over the next three months, organized as Now / Next / Later. No fake deadlines — the order reflects what unblocks the most users, in the rough sequence it will land. Quick links: - Rendered roadmap page: https://hydron.app/roadmap - llms.txt directive: https://hydron.app/llms.txt - Documentation: https://hydron.app/docs - Blog: https://hydron.app/blog ## Now _Actively underway, shipping in the early weeks of the period._ ### Zero-downtime rolling deploys - **Theme**: Reliability - **Anchor**: https://hydron.app/roadmap#zero-downtime-deploys Every Hydron deploy will swap containers in place — the new version goes live and proves itself healthy before the old one shuts down. No more 502s during release windows, no more reload-and-pray. **Why we're building it:** If your users see a 502 every time you ship, you aren't really shipping continuously. Zero-downtime is table stakes for production. **Best for:** Customer-facing applications, High-traffic APIs, Teams that deploy multiple times a day. ## Next _Starts once the Now items land — expected mid-period._ ### Team accounts with shared billing & roles - **Theme**: Teams & collaboration - **Anchor**: https://hydron.app/roadmap#team-accounts Multiple developers, one billing account, the right permissions per person. Invite your team, set roles, and stop juggling shared logins. **Why we're building it:** Hydron stops being a tool and starts being a platform the moment a teammate wants in. Teams should never have to share passwords, and your bill shouldn't depend on which person clicked Deploy. **Best for:** Teams of 3–15 developers, Agencies managing client deployments, Anyone who wants a deploy log instead of a Slack thread. ### MCP server — deeper agent control over Hydron - **Theme**: AI agent surface - **Anchor**: https://hydron.app/roadmap#mcp-server Expand the Model Context Protocol server so any AI coding agent — Claude Code, Cursor, Codex, anything that speaks MCP — can drive a complete Hydron deployment end-to-end without ever leaving the editor. **Why we're building it:** If an AI agent is doing the building, it should also be able to do the shipping. Hydron's MCP server is the bridge — making sure no operation requires breaking out of the AI workflow you're already in. **Best for:** Claude Code, Cursor, Codex & Aider users, Teams running agentic dev loops, Anyone building on top of the Hydron API. ## Later _Larger investments lined up for the back end of the period._ ### Vercel / Heroku / Render migration assistant - **Theme**: Migration & onboarding - **Anchor**: https://hydron.app/roadmap#migration-assistant Point Hydron at your current host. We'll read your build settings, environment variables, and add-ons (Postgres, Redis, ...), generate the matching Hydron configuration, and walk you through database transfer and DNS cutover. **Why we're building it:** Migrating off a managed platform usually takes days of manual config rebuilding. Most of that work is mechanical — exactly the kind of thing AI should do for you. **Best for:** Vercel, Heroku, Render, and Railway customers, Teams whose bill outgrew the convenience, Anyone who wants out without a re-platforming project. ## How we plan - **Production trust before new surface area** — reliability work comes ahead of net-new product. - **Honest pricing, in public** — if you can't price Hydron without talking to us, that's a bug. - **Teams & agencies are first-class users** — multi-user accounts are a default, not an enterprise tier. - **Migration is the unlock** — the cost of getting in should never outweigh the cost of staying put. ## Requesting a feature Feature requests go through the contact form at https://hydron.app/contact. Specific, concrete asks tied to a real workflow carry the most weight.