### Install Dependencies using Pip

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Installs project dependencies using 'pip'. This command installs the project in editable mode.

```bash
pip install -e .
```

--------------------------------

### Install Dependencies using UV

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Installs project dependencies using the 'uv' package manager. 'uv sync' is recommended for efficient dependency management.

```bash
uv sync
```

--------------------------------

### Run FastAPI Application

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Starts the FastAPI development server using 'uvicorn'. The '--reload' flag enables automatic server restarts on code changes. The API will be accessible at http://localhost:8000.

```bash
uvicorn app.main:app --reload
```

--------------------------------

### Clone Repository using Git

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Clones the project repository from GitHub. This is a prerequisite step before installing dependencies and running the application.

```bash
git clone https://github.com/e2b-dev/claude-code-fastapi
cd claude-code-fastapi
```

--------------------------------

### Access API Documentation

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Navigates to the interactive API documentation for the running FastAPI application. This typically uses Swagger UI and is available at the '/docs' endpoint.

```bash
http://localhost:8000/docs
```

--------------------------------

### Start New Chat Session with GitHub Repo

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Initiates a new chat session with Claude Code via a POST request to the '/chat' endpoint. It includes a prompt and an optional GitHub repository URL to clone into the sandbox.

```bash
curl -X POST "http://localhost:8000/chat" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Can you add GPT 3.5-Turbo to the list of models and open a pull request on GitHub",
    "repo": "https://github.com/e2b-dev/fragments"
  }'
```

--------------------------------

### Start or Resume Chat Session

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Initiates a new chat session with Claude Code or continues an existing one using a session ID. For new sessions, a GitHub repository can be provided to clone.

```APIDOC
## POST /chat

Start a new conversation with Claude Code or resume an existing one.

### Method
POST

### Endpoint
`/chat` or `/chat/{session}`

### Parameters
#### Path Parameters
- **session** (string) - Optional - The ID of an existing session to resume.

#### Query Parameters
None

#### Request Body
- **prompt** (string) - Required - The prompt or command to send to Claude Code.
- **repo** (string) - Optional - The URL of a GitHub repository to clone. This is only used when starting a new session (i.e., not providing a `session` path parameter).

### Request Example
```json
{
  "prompt": "Can you add GPT 3.5-Turbo to the list of models and open a pull request on GitHub",
  "repo": "https://github.com/e2b-dev/fragments"
}
```

### Response
#### Success Response (200)
- **type** (string) - Type of the response, typically 'result'.
- **subtype** (string) - Subtype of the result, typically 'success'.
- **is_error** (boolean) - Indicates if an error occurred, false for success.
- **duration_ms** (integer) - Total duration of the operation in milliseconds.
- **duration_api_ms** (integer) - Duration of the API call processing in milliseconds.
- **num_turns** (integer) - Number of conversational turns.
- **result** (string) - The output or result from Claude Code.
- **session_id** (string) - The ID of the current session. This is returned for both new and resumed sessions.
- **total_cost_usd** (number) - The total cost of the operation in USD.
- **usage** (object) - Token usage details.
  - **input_tokens** (integer) - Number of input tokens.
  - **cache_creation_input_tokens** (integer) - Input tokens used for cache creation.
  - **cache_read_input_tokens** (integer) - Input tokens read from cache.
  - **output_tokens** (integer) - Number of output tokens.
  - **server_tool_use** (object) - Tool usage details.
    - **web_search_requests** (integer) - Number of web search requests made.
  - **service_tier** (string) - The service tier used.
- **permission_denials** (array) - List of permission denials, if any.

#### Response Example
```json
{
  "type": "result",
  "subtype": "success",
  "is_error": false,
  "duration_ms": 216401,
  "duration_api_ms": 216234,
  "num_turns": 81,
  "result": "Perfect! I have successfully completed all the tasks:\n\n✅ **All tasks completed!** The pull request has been created at: https://github.com/e2b-dev/fragments/pull/170\n\n## Summary\n\nI've successfully added GPT-3.5 Turbo to the model list and opened a pull request with the following changes:\n\n- **Added GPT-3.5 Turbo** to `/home/user/fragments/lib/models.json:108-114`\n- **Model configuration**:\n  - ID: `gpt-3.5-turbo`\n  - Provider: OpenAI\n  - Name: GPT-3.5 Turbo  \n  - Multimodal: false\n- **Validated** JSON syntax is correct\n- **Created branch** `add-gpt-3.5-turbo` \n- **Opened PR #170** with comprehensive description\n\nThe pull request is ready for review and includes all the necessary changes to integrate GPT-3.5 Turbo into the fragments application.",
  "session_id": "038b769b-4717-47ca-be02-2a49bd7da978",
  "total_cost_usd": 1.1459241,
  "usage": {
    "input_tokens": 300,
    "cache_creation_input_tokens": 77458,
    "cache_read_input_tokens": 2087724,
    "output_tokens": 13935,
    "server_tool_use": {
      "web_search_requests": 0
    },
    "service_tier": "standard"
  },
  "permission_denials": []
}
```
```

--------------------------------

### Configure Environment Variables

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Sets up required and optional environment variables in a `.env` file for application configuration, including API keys for Anthropic, E2B, and optionally GitHub and Context7.

```env
# Required
ANTHROPIC_API_KEY=your_anthropic_api_key_here
E2B_API_KEY=your_e2b_api_key_here

# Optional
E2B_SANDBOX_TEMPLATE=claude-code-dev  # or claude-code for production
GITHUB_PAT=your_github_personal_access_token_here
CONTEXT7_API_KEY=your_context7_api_key_here  # for MCP Context7 server
```

--------------------------------

### Build Development Sandbox Template

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Builds the development E2B sandbox template using a Python script. This template is configured with specific CPU and memory resources and assigned an alias.

```bash
cd template
python build_dev.py
```

--------------------------------

### API Response Example

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

An example of a successful response from the chat API endpoint, detailing task completion, cost, usage statistics, and potential permission denials.

```json
{
  "type": "result",
  "subtype": "success",
  "is_error": false,
  "duration_ms": 216401,
  "duration_api_ms": 216234,
  "num_turns": 81,
  "result": "Perfect! I have successfully completed all the tasks:\n\n✅ **All tasks completed!** The pull request has been created at: https://github.com/e2b-dev/fragments/pull/170\n\n## Summary\n\nI've successfully added GPT-3.5 Turbo to the model list and opened a pull request with the following changes:\n\n- **Added GPT-3.5 Turbo** to `/home/user/fragments/lib/models.json:108-114`\n- **Model configuration**:\n  - ID: `gpt-3.5-turbo`\n  - Provider: OpenAI\n  - Name: GPT-3.5 Turbo  \n  - Multimodal: false\n- **Validated** JSON syntax is correct\n- **Created branch** `add-gpt-3.5-turbo` \n- **Opened PR #170** with comprehensive description\n\nThe pull request is ready for review and includes all the necessary changes to integrate GPT-3.5 Turbo into the fragments application.",
  "session_id": "038b769b-4717-47ca-be02-2a49bd7da978",
  "total_cost_usd": 1.1459241,
  "usage": {
    "input_tokens": 300,
    "cache_creation_input_tokens": 77458,
    "cache_read_input_tokens": 2087724,
    "output_tokens": 13935,
    "server_tool_use": {
      "web_search_requests": 0
    },
    "service_tier": "standard"
  },
  "permission_denials": []
}
```

--------------------------------

### Build Production Sandbox Template

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Builds the production E2B sandbox template using a Python script. This template is configured with specific CPU and memory resources and assigned an alias.

```bash
cd template
python build.py
```

--------------------------------

### Resume Existing Chat Session

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Resumes an existing conversation using a session ID by sending a POST request to the '/chat/{session}' endpoint. The prompt is provided in the request body.

```bash
curl -X POST "http://localhost:8000/chat/038b769b-4717-47ca-be02-2a49bd7da978" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Now modify the script to accept a name parameter"
  }'
```

--------------------------------

### Configure MCP Servers with JSON

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

This JSON configuration allows you to add custom MCP servers to Claude Code. It specifies the server type, URL, and necessary headers like authentication tokens. Ensure the CONTEXT7_API_KEY environment variable is set for authorization.

```json
{
  "mcpServers": {
    "context7": {
      "type": "http",
      "url": "https://mcp.context7.com/mcp",
      "headers": {
        "Authorization": "Bearer ${CONTEXT7_API_KEY}"
      }
    }
  }
}
```

--------------------------------

### API Request Body Schema

Source: https://github.com/e2b-dev/claude-code-fastapi/blob/main/README.md

Defines the structure of the JSON request body for the /chat and /chat/{session} API endpoints. It includes the 'prompt' (required) and 'repo' (optional, only for new sessions).

```json
{
  "prompt": "string", // Required: The prompt/command for Claude Code
  "repo": "string" // Optional: GitHub repository URL to clone (only for new sessions)
}
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.