Coding agents
Connect your coding agent to your database.
If you build with an AI coding agent, it can use your CoyoteDB database like any Postgres database. Give it a normal connection string and the short note below.
Three steps
A connection string and a note.
- Print the connection details. On a node, run
coyote conninfo. It prints the connection string, with the password. - Put it in an environment variable. For example
DATABASE_URL, so the password never sits in your code. - Tell your agent. Paste the note below into your agent's instructions. It works the same for Claude Code, Codex, Grok Build and other agents.
# on a node
$ coyote conninfo
Tell your AI agent
Paste this into your agent's instructions.
Any agent that can use a Postgres library can work with CoyoteDB the same way as with any Postgres database. This note tells it how, and what to ask you.
This project's database is CoyoteDB: real PostgreSQL 18 running on one or more of my own computers ("nodes"), kept in sync.
How to connect
- Treat it as a normal PostgreSQL database. Use a standard postgresql:// connection string and ordinary Postgres libraries, ORMs and migrations.
- Point the connection at the address I give you. Never hard-code which node is "main": it can change.
- Read the connection string from an environment variable (for example DATABASE_URL); never put the password in code.
Ask me for
- the host name or IP address of a node (or of each node),
- the port, the database name and the user,
- how the password is provided to the app.
Keep in mind
- If a computer stops, writes can pause for about 30 seconds while the nodes sort it out; use connection retries.
- Copies between nodes are not backups: a destructive migration reaches every node. Ask before dropping or rewriting data.