> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gentic.chat/llms.txt
> Use this file to discover all available pages before exploring further.

# Gentic Quickstart: From Sign-Up to First Pull Request

> Create a Gentic project, deploy the agent worker, write your first issue, and receive an AI-generated pull request in a single sitting.

This guide takes you from a blank account to a merged pull request. You will sign up, connect a GitHub repository, deploy the agent worker on a server, create an issue, and watch the AI agent write code and open a PR — all in one sitting.

<Steps>
  ### Sign up at gentic.chat

  Open [gentic.chat](https://gentic.chat) in your browser and create a free account. After email confirmation, you land on the main dashboard where your projects and issues live.

  ### Add a project

  Navigate to **Settings** in the top navigation, then find the **Projects** section. Click **Add project** and enter your repository in `owner/repo` format — for example, `acme-corp/backend`. Gentic stores this slug and the agent worker uses it to construct the full Git remote URL when it clones your code.

  <Tip>
    You can add as many projects as you need. Each issue is scoped to a single project, so the worker always knows which repository to clone.
  </Tip>

  ### Connect GitHub

  Still in **Settings**, open the **GitHub integration** section and follow the OAuth flow to authorize Gentic. This grants the platform permission to open pull requests and manage branches on your behalf. Make sure you grant access to the repository you added in the previous step.

  ### Install and configure the agent worker

  The agent worker is a CLI you run on a server, VM, or any machine that can reach both the Gentic API and GitHub. Download the latest pre-built binary for your platform from the [GitHub Releases page](https://github.com/kprovorov/gentic/releases), or build from source.

  Once the binary is on your server, authenticate it with your Gentic API key:

  ```bash theme={null}
  gentic auth login
  ```

  The interactive prompt asks for the API URL and your key. For non-interactive or scripted setup, pass both flags directly:

  ```bash theme={null}
  gentic auth login \
    --api-url https://gentic.chat/api/v1 \
    --api-key your-api-key-here
  ```

  Alternatively, create a `.env` file next to the binary and set the variables there:

  ```bash theme={null}
  GENTIC_API_URL=https://gentic.chat/api/v1
  GENTIC_API_KEY=your-api-key-here
  GIT_REMOTE_BASE=git@github.com:
  POLL_INTERVAL_MS=3000
  ```

  Then install and start the worker as a managed OS service so it restarts automatically:

  ```bash theme={null}
  gentic start
  ```

  Verify the worker is authenticated and running:

  ```bash theme={null}
  gentic auth status
  ```

  <Note>
    The agent worker must be running before you move an issue to Todo. If no worker is polling, queued issues stay queued indefinitely. See the [Agent Worker installation guide](/agent/overview) for full details on service management, SSH deploy keys, and per-agent prerequisites.
  </Note>

  ### Create your first issue

  Back in the Gentic web app, open your project and click **New issue**. Fill in the fields:

  | Field      | Value                                                             |
  | ---------- | ----------------------------------------------------------------- |
  | **Title**  | A short, descriptive name for the task                            |
  | **Type**   | `Feature` (or `Bug`, `Chore`, etc.)                               |
  | **Agent**  | `Claude Code` or `Codex`                                          |
  | **Prompt** | A detailed description of exactly what the agent should implement |
  | **Status** | `Draft`                                                           |

  Save the issue. Draft issues are never picked up by the worker, so you can take your time refining the prompt and attaching any supporting files before you commit to running it.

  <Tip>
    Write prompts the same way you would write a detailed ticket for a junior engineer: describe the goal, the acceptance criteria, any relevant file paths, and the coding style you expect. The more context you give, the better the output.
  </Tip>

  ### Move the issue to Todo

  When your issue is ready, change its **Status** from `Draft` to `Todo`. This single action queues the issue for the agent worker — Gentic sets `run_status = queued` and records your prompt as the first message in the issue's conversation thread.

  On the worker's next poll (within a few seconds), it claims the issue, clones the repository into a fresh working directory, and starts the AI agent.

  ### Watch the pull request appear

  Switch to the **Issues** view and open your issue. You will see `run_status` move through `queued → running` as the worker picks it up, and real-time agent output stream into the transcript panel. When the agent finishes, it commits its changes and opens a pull request on GitHub. The issue status updates to `Done` and a link to the PR appears in the issue.

  Open the pull request on GitHub, review the changes, and merge when you are satisfied.

  <CardGroup cols={2}>
    <Card title="Agent Worker Setup" icon="server" href="/agent/overview">
      Full installation guide: service management, SSH keys, Claude Code and Codex credentials.
    </Card>

    <Card title="How It Works" icon="circle-info" href="/how-it-works">
      A deeper look at the end-to-end flow from issue creation to pull request.
    </Card>
  </CardGroup>
</Steps>