Roadmap
Boost.space is the synced data layer where humans and AI agents run a business together. Everything below serves one arc: your business data becoming something an agent can understand, act on, and be trusted with.
In development
Your own Python as an automation step, in a secure sandbox. Releasing shortly.
Describe it in chat; an agent writes that Python, tests it, and ships it live.
Batch operations, an impact preview before anything destructive runs, leaner tool metadata.
More of the system it can operate, and more work it finishes end to end.
Agents learn what a field means, not only what it contains.
Custom modules an order of magnitude larger, without losing their speed.
The four-language family completes, from the same generated pipeline.
Next up
The shape of your whole system — and what an agent sees when it looks at it.
An agent as a member of your space, with its own permissions and audit trail.
Find the record you meant, not the one that happened to match a keyword.
Fewer concepts to learn, less to configure, a shorter path to a system that works.
Exploring
Say what must stay true — never oversell, one clean record per customer — and get proof it stayed true.
Small writes flow; big or irreversible ones wait with a dry-run diff and their blast radius.
A proven schema with its automations and views, fitted by an agent to the data you already have.
The dashboard you asked for, wired to your data and living inside your permissions.
Answer a question about your data once, and it becomes a rule you are never asked about again.
Shipped recently
| Feature | What it does |
|---|---|
| More capable MCP tools | Connected agents now build structure too — modules, spaces, field groups, fields — and read pre-computed aggregations instead of raw history |
| Agent Chat | The built-in agent: ask, act, and build automations in plain language |
| Vector database | Spaces embedded for meaning, so retrieval works by intent |
| Build a module with AI | Describe what you want to track; AI generates the module and its typed fields |
| TypeScript and Python SDKs | Typed clients generated from the REST API spec |
| Formula and button fields | Values computed from other fields, and an action button on the record |
| Docs built for agents | Raw markdown per page, an llms.txt index, and a docs MCP endpoint |
The full record lives in the API changelog for every public API change, the SDK changelog for client releases, and the release notes for product announcements.
Direction, not delivery dates — priorities follow what customers ask for. Waiting on something above? Tell us.
