Jira

PerSQL gives each Jira issue its own isolated SQLite branch database — the scratch space an agent needs to work a ticket. The branch spawns when the issue is created and is cleaned up when it transitions to done.

Add the integration

1. Go to jira.persql.com/install/start, authorize via Atlassian, and connect your site to your PerSQL workspace.

2. From then on, every new issue gets a schema-only branch database named issue-<key> (for example issue-PROJ-318).

A database per issue

Point a coding agent at the Jira MCP for the per-issue recipe:

https://jira.persql.com/mcp

Authenticate with a PerSQL bearer token. issue_recipe returns the snippet that spawns or resets the issue’s branch; issue_list shows the active ones. The agent reads and writes that branch as it works the ticket:

import { PerSQL } from "@persql/sdk";

const db = new PerSQL({ token: process.env.PERSQL_TOKEN })
  .database("<namespace>/issue-PROJ-318");

await db.query("SELECT count(*) AS n FROM sqlite_master");

When the issue transitions to done, the branch is dropped.