Connect your coding agent
How do I point Claude Code, Cursor, or Codex at a Backenly project?
Read
Documentation
Seven guides covering the whole path: connecting a coding agent over MCP, the build loop and the checks that run after it, the data API your frontend calls, the authorization model, what the autonomy loop does once you are live, and how to run all of it yourself.
2 guides
Connect an agent, build the first backend, and wire a frontend to it.
5 guides
The architecture, the authorization model, what happens after launch, and how to run the whole thing yourself.
What exactly can my frontend call, and how does it authenticate?
Read
What actually happens between my agent and my database?
Read
Who can read which rows, and how do I know the rule actually holds?
Read
Who is watching this backend when I am not, and what will it do without asking?
Read
What do I actually run, and what do I take on by running it?
Read
Elsewhere
Some of what you might want is better read at its source than paraphrased here.
The complete machine-readable reference: every endpoint, the full tool table, plan limits, and the architecture. Agents can also fetch it at run time with the fetch_docs tool.
The platform is Apache-2.0 and the client libraries are MIT. Everything the hosted product runs is in the repository, including the autonomy engine.
Capacity per plan: projects, monthly active users, storage, AI credits, and which features are plan-gated. Kept in one place so no guide restates a number that can move.
Five workflows with the problem, what Backenly does, what stays yours, and the known limitations of each.
Explore Backenly: