> For the complete documentation index, see [llms.txt](https://burp-ai-agent.six2dez.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://burp-ai-agent.six2dez.com/developer/adding-backend.md).

# Adding a Backend

The extension uses Java's `ServiceLoader` mechanism for backend discovery, making it straightforward to add new AI backends without modifying core code.

## Steps

### 1. Implement the Backend Adapter

Implement a factory, a backend, and a connection:

* **`AiBackendFactory`**: zero-argument ServiceLoader factory whose `create()` returns the backend.
* **`AiBackend`**: owns `id`, `displayName`, availability/health checks, and `launch(config)`.
* **`AgentConnection`**: asynchronous send/stream/stop contract returned by `launch`.

CLI connections can also implement `DiagnosableConnection` to expose exit codes and output tails. HTTP connections commonly implement `UsageAwareConnection` and, when supported, `JsonModeCapable`.

### 2. Register via ServiceLoader

Create a file at:

```
META-INF/services/com.six2dez.burp.aiagent.backends.AiBackendFactory
```

Add the fully qualified class name of your factory:

```
com.example.mybackend.MyBackendFactory
```

### 3. Add UI Configuration Fields (Optional)

If your backend needs custom settings (API URL, model name, etc.), add configuration keys to `AgentSettings.kt`, wire load/save in `SettingsPanelSettingsIO.kt`, and place the controls in the appropriate settings panel. If persisted shape changes, add a forward-only schema step as described in [Settings Migration](/developer/settings-migration.md).

### 4. Ensure Audit Logging

Use the backend ID consistently so audit logs can trace which backend was used for each interaction.

### 5. Package and Deploy

**Option A: Built-in** Add your backend to the main source tree and rebuild the extension JAR.

**Option B: Drop-in JAR** Package your backend as a standalone JAR and place it in `~/.burp-ai-agent/backends/`. The `BackendRegistry` will discover it automatically via `URLClassLoader` on startup.

## Backend Interface

```kotlin
class MyBackendFactory : AiBackendFactory {
    override fun create(): AiBackend = MyBackend()
}

class MyBackend : AiBackend {
    override val id = "my-backend"
    override val displayName = "My Backend"
    override val supportsSystemRole = true

    override fun launch(config: BackendLaunchConfig): AgentConnection =
        MyConnection(config)

    override fun isAvailable(settings: AgentSettings): Boolean = true

    override fun healthCheck(settings: AgentSettings): HealthCheckResult =
        HealthCheckResult.Healthy
}

class MyConnection(
    private val config: BackendLaunchConfig,
) : AgentConnection {
    override fun isAlive(): Boolean = true

    override fun send(
        text: String,
        history: List<ChatMessage>?,
        onChunk: (String) -> Unit,
        onComplete: (Throwable?) -> Unit,
        systemPrompt: String?,
        jsonMode: Boolean,
        maxOutputTokens: Int?,
    ) {
        // Run network/process work off the Swing EDT, call onChunk as data arrives,
        // then call onComplete exactly once.
    }

    override fun stop() {
        // Cancel I/O and release process/network resources.
    }
}
```

`BackendLaunchConfig` carries the resolved command or base URL, model, headers, timeouts, session IDs, determinism flag, context window, and shared Montoya HTTP transport. Implement `DiagnosableConnection`, `UsageAwareConnection`, `SessionAwareConnection`, `JsonModeCapable`, or `supportsSystemRole` only when the backend genuinely supports that capability.

## Tips

* Follow existing backend implementations as examples: `OllamaBackendFactory` for HTTP and `CodexCliBackendFactory` or `CopilotCliBackendFactory` for CLI.
* Use the shared Montoya HTTP transport and retry/circuit-breaker support instead of creating an unrelated client stack.
* Never block the Swing EDT. The connection callback contract is asynchronous.
* Ensure `onComplete` fires exactly once on success, error, cancellation, and timeout.
* CLI backends must manage stdin/stdout, exit status, temporary files, cancellation, and Windows `.cmd` shims.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://burp-ai-agent.six2dez.com/developer/adding-backend.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.