Remove A2A from haiku.rag, turn it into an example

CHANGELOG.md

@@ -5,12 +5,13 @@
 
 - **Configuration**: All CLI commands now properly support `--config` parameter for specifying custom configuration files
 - Configuration loading consolidated across CLI, app, and client with consistent resolution order
-- `HaikuRAGApp`, MCP server, and A2A server now accept `config` parameter for programmatic configuration
+- `HaikuRAGApp` and MCP server now accept `config` parameter for programmatic configuration
 - Updated CLI documentation to clarify global vs per-command options
 - **BREAKING**: Standardized configuration filename to `haiku.rag.yaml` in user directories (was incorrectly using `config.yaml`). Users with existing `config.yaml` in their user directory will need to rename it to `haiku.rag.yaml`
 
 ### Removed
 
+- **BREAKING**: A2A (Agent-to-Agent) protocol support has been moved to a separate self-contained package in `examples/a2a-server/`. The A2A server is no longer part of the main haiku.rag package. Users who need A2A functionality can install and run it from the examples directory with `cd examples/a2a-server && uv sync`.
 - **BREAKING**: Removed deprecated `.env`-based configuration system. The `haiku-rag init-config --from-env` command and `load_config_from_env()` function have been removed. All configuration must now be done via YAML files. Environment variables for API keys (e.g., `OPENAI_API_KEY`, `ANTHROPIC_API_KEY`) and service URLs (e.g., `OLLAMA_BASE_URL`) are still supported and can be set via `.env` files.
 
 ## [0.14.1] - 2025-11-06

README.md

@@ -16,7 +16,6 @@ Retrieval-Augmented Generation (RAG) library built on LanceDB.
 - **File monitoring**: Auto-index files when run as server
 - **40+ file formats**: PDF, DOCX, HTML, Markdown, code files, URLs
 - **MCP server**: Expose as tools for AI assistants
-- **A2A agent**: Conversational agent with context and multi-turn dialogue
 - **CLI & Python API**: Use from command line or Python
 
 ## Installation
@@ -29,7 +28,7 @@ Retrieval-Augmented Generation (RAG) library built on LanceDB.
 uv pip install haiku.rag
 ```
 
-Includes all features: document processing, all embedding providers, rerankers, and A2A agent support.
+Includes all features: document processing, all embedding providers, and rerankers.
 
 ### Slim Package (Minimal Dependencies)
 
@@ -148,32 +147,13 @@ haiku-rag serve --stdio
 
 Provides tools for document management and search directly in your AI assistant.
 
-## A2A Agent
-
-Run as a conversational agent with the Agent-to-Agent protocol:
-
-```bash
-# Start the A2A server
-haiku-rag serve --a2a
-
-# Connect with the interactive client (in another terminal)
-haiku-rag a2aclient
-```
-
-The A2A agent provides:
-
-- Multi-turn dialogue with context
-- Intelligent multi-search for complex questions
-- Source citations with titles and URIs
-- Full document retrieval on request
-
 ## Examples
 
 See the [examples directory](examples/) for working examples:
 
 - **[Interactive Research Assistant](examples/ag-ui-research/)** - Full-stack research assistant with Pydantic AI and AG-UI featuring human-in-the-loop approval and real-time state synchronization
-- **[Docker Setup](examples/docker/)** - Complete Docker deployment with file monitoring, MCP server, and A2A agent
+- **[Docker Setup](examples/docker/)** - Complete Docker deployment with file monitoring and MCP server
-- **[A2A Security](examples/a2a-security/)** - Authentication examples (API key, OAuth2, GitHub)
+- **[A2A Server](examples/a2a-server/)** - Self-contained A2A protocol server package with conversational agent interface
 
 ## Documentation
 
@@ -185,7 +165,6 @@ Full documentation at: https://ggozad.github.io/haiku.rag/
 - [Python API](https://ggozad.github.io/haiku.rag/python/) - Complete API docs
 - [Agents](https://ggozad.github.io/haiku.rag/agents/) - QA agent and multi-agent research
 - [MCP Server](https://ggozad.github.io/haiku.rag/mcp/) - Model Context Protocol integration
-- [A2A Agent](https://ggozad.github.io/haiku.rag/a2a/) - Agent-to-Agent protocol support
 - [Benchmarks](https://ggozad.github.io/haiku.rag/benchmarks/) - Performance Benchmarks
 
 mcp-name: io.github.ggozad/haiku-rag

docker/Dockerfile

@@ -33,8 +33,8 @@ ENV DEFAULT_DATA_DIR=/data
 
 ENV PATH="/app/.venv/bin:$PATH"
 
-# Expose ports for MCP and A2A
+# Expose port for MCP server
-EXPOSE 8000 8001
+EXPOSE 8001
 
-# Run all services (monitoring, MCP, A2A)
+# Run all services (monitoring, MCP)
-CMD ["python", "-m", "haiku.rag.cli", "serve", "--monitor", "--mcp", "--mcp-port", "8001", "--a2a", "--a2a-host", "0.0.0.0", "--a2a-port", "8000", "--db", "/data/haiku.rag.lancedb"]
+CMD ["python", "-m", "haiku.rag.cli", "serve", "--monitor", "--mcp", "--mcp-port", "8001", "--db", "/data/haiku.rag.lancedb"]

docker/README.md

@@ -1,6 +1,6 @@
 # haiku.rag Docker Image
 
-Pre-built images are available at `ghcr.io/ggozad/haiku.rag` with all extras (voyageai, mxbai, a2a).
+Pre-built images are available at `ghcr.io/ggozad/haiku.rag` with all extras (voyageai, mxbai).
 
 ## Using Pre-built Image
 
@@ -33,7 +33,7 @@ See [Configuration docs](https://ggozad.github.io/haiku.rag/configuration/) for
 Mount your config file and data directory:
 
 ```bash
-docker run -p 8000:8000 -p 8001:8001 \
+docker run -p 8001:8001 \
   -v $(pwd)/haiku.rag.yaml:/app/haiku.rag.yaml \
   -v $(pwd)/data:/data \
   ghcr.io/ggozad/haiku.rag:latest
@@ -44,7 +44,7 @@ The container will automatically use the mounted `haiku.rag.yaml` configuration
 For API keys (OpenAI, Anthropic, etc.), pass them as environment variables:
 
 ```bash
-docker run -p 8000:8000 -p 8001:8001 \
+docker run -p 8001:8001 \
   -v $(pwd)/haiku.rag.yaml:/app/haiku.rag.yaml \
   -v $(pwd)/data:/data \
   -e OPENAI_API_KEY=your-key-here \

docs/cli.md

@@ -171,41 +171,18 @@ haiku-rag serve --mcp
 # MCP server (stdio transport)
 haiku-rag serve --mcp --stdio
 
-# A2A server only
-haiku-rag serve --a2a
-
 # File monitoring only
 haiku-rag serve --monitor
 
-# All services
+# Both services
-haiku-rag serve --monitor --mcp --a2a
+haiku-rag serve --monitor --mcp
 
-# Custom ports
+# Custom port
-haiku-rag serve --mcp --mcp-port 9000 --a2a --a2a-port 9001
+haiku-rag serve --mcp --mcp-port 9000
 ```
 
 See [Server Mode](server.md) for details on available services.
 
-### A2A Interactive Client
-
-Connect to and chat with haiku.rag's A2A server:
-
-```bash
-# Connect to local server
-haiku-rag a2aclient
-
-# Connect to remote server
-haiku-rag a2aclient --url https://example.com:8000
-```
-
-The interactive client provides:
-- Rich markdown rendering of agent responses
-- Multi-turn conversation with context
-- Agent card discovery and display
-- Compact artifact summaries
-
-See [A2A documentation](a2a.md) for more details.
-
 ## Settings
 
 View current configuration settings:

docs/configuration.md

@@ -98,9 +98,6 @@ providers:
     rerank_base_url: ""
     qa_base_url: ""
     research_base_url: ""
-
-a2a:
-  max_contexts: 1000
 ```
 
 ## Programmatic Configuration

docs/index.md

@@ -12,7 +12,6 @@
 - **File monitoring**: Automatically index files when run as a server
 - **Extended file format support**: Parse 40+ file formats including PDF, DOCX, HTML, Markdown, code files and more. Or add a URL!
 - **MCP server**: Exposes functionality as MCP tools
-- **A2A agent**: Conversational agent with context and multi-turn dialogue support
 - **CLI commands**: Access all functionality from your terminal
   - Add sources from text, files, or URLs, optionally with a human‑readable title
 - **Python client**: Call `haiku.rag` from your own python applications
@@ -59,7 +58,6 @@ haiku-rag ask "Who is the author of haiku.rag?"
 - [CLI](cli.md) - Command line interface usage
 - [Server](server.md) - File monitoring and server mode
 - [MCP](mcp.md) - Model Context Protocol integration
-- [A2A](a2a.md) - Agent-to-Agent conversational protocol
 - [Python](python.md) - Python API reference
 - [Agents](agents.md) - QA agent and multi-agent research
 

docs/installation.md

@@ -14,7 +14,6 @@ The full package includes **all features and extras**:
 - **Document processing** (Docling) - PDF, DOCX, PPTX, images, and 40+ file formats
 - **All embedding providers** - VoyageAI
 - **All rerankers** - MixedBread AI, Cohere, Zero Entropy
-- **A2A agent** - Agent-to-Agent protocol support
 
 This is the easiest way to get started with all features enabled.
 
@@ -36,7 +35,6 @@ The slim package has minimal dependencies and lets you install only what you nee
 - `docling` - PDF, DOCX, PPTX, images, and other document formats
 - `voyageai` - VoyageAI embeddings
 - `mxbai` - MixedBread AI reranking
-- `a2a` - Agent-to-Agent protocol support
 - `cohere` - Cohere reranking
 - `zeroentropy` - Zero Entropy reranking
 
@@ -74,4 +72,4 @@ Run the container with all services:
 docker run -p 8000:8000 -p 8001:8001 -v $(pwd)/data:/data ghcr.io/ggozad/haiku.rag:latest
 ```
 
-This starts the MCP server on port 8001 and A2A server on port 8000, with data persisted to `./data`.
+This starts the MCP server on port 8001, with data persisted to `./data`.

docs/server.md

@@ -1,10 +1,10 @@
 # Server Mode
 
-The server provides automatic file monitoring, MCP functionality, and A2A agent support.
+The server provides automatic file monitoring and MCP functionality.
 
 ## Starting the Server
 
-The `serve` command requires at least one service flag. You can enable file monitoring, MCP server, A2A server, or any combination:
+The `serve` command requires at least one service flag. You can enable file monitoring, MCP server, or both:
 
 ### MCP Server Only
 
@@ -17,31 +17,19 @@ Transport options:
 - `--stdio` - Standard input/output transport
 - `--mcp-port` - Custom port (default: 8001)
 
-### A2A Server Only
-
-```bash
-haiku-rag serve --a2a
-```
-
-Options:
-- `--a2a-host` - Host to bind to (default: 127.0.0.1)
-- `--a2a-port` - Port to bind to (default: 8000)
-
-See [A2A documentation](a2a.md) for details on the conversational agent.
-
 ### File Monitoring Only
 
 ```bash
 haiku-rag serve --monitor
 ```
 
-### All Services
+### Both Services
 
 ```bash
-haiku-rag serve --monitor --mcp --a2a
+haiku-rag serve --monitor --mcp
 ```
 
-This will start file monitoring, MCP server on port 8001, and A2A server on port 8000.
+This will start file monitoring and MCP server on port 8001.
 
 ## File Monitoring
 

examples/README.md

@@ -22,44 +22,25 @@ See `ag-ui-research/README.md` for setup instructions.
 Complete Docker setup for running haiku.rag with all services:
 - File monitoring for automatic document indexing
 - MCP server for AI assistant integration
-- A2A agent for conversational interactions
 
 See `docker/README.md` for setup instructions.
 
-## A2A Security Examples
+## A2A Server
 
-**Directory:** `a2a-security/`
+**Directory:** `a2a-server/`
 
-Three examples showing how to add authentication to haiku.rag's A2A server:
+Self-contained A2A (Agent-to-Agent) protocol server package that provides a conversational agent interface with its own CLI and dependencies.
 
-### API Key Authentication
+Features:
+- Conversational context with multi-turn dialogue support
+- Interactive CLI client for testing
+- Security examples (API key, OAuth2 with GitHub, enterprise OAuth2)
+- Full documentation and installation instructions
 
-**File:** `a2a-security/apikey_example.py`
+See `a2a-server/README.md` for complete setup and usage instructions.
-
-Simple header-based authentication suitable for internal services and development.
 
+Install locally:
 ```bash
-python examples/a2a-security/apikey_example.py /path/to/database.lancedb
+cd a2a-server
+uv sync
 ```
-
-### OAuth2 GitHub Authentication
-
-**File:** `a2a-security/oauth2_github.py`
-
-GitHub Personal Access Token authentication for GitHub-integrated services.
-
-```bash
-python examples/a2a-security/oauth2_github.py /path/to/database.lancedb
-```
-
-### OAuth2 Enterprise Authentication
-
-**File:** `a2a-security/oauth2_example.py`
-
-Full OAuth2 with JWT verification for enterprise environments.
-
-```bash
-python examples/a2a-security/oauth2_example.py /path/to/database.lancedb
-```
-
-See individual files for detailed setup instructions and usage examples.

examples/a2a-server/README.md

@@ -0,0 +1,117 @@
+# haiku-rag-a2a
+
+A2A (Agent-to-Agent) protocol server for haiku.rag. This package provides a conversational agent interface that maintains conversation history and context across multiple turns.
+
+## Features
+
+- **Conversational Context**: Maintains full conversation history including tool calls and results
+- **Multi-turn Dialogue**: Supports follow-up questions with pronoun resolution ("he", "it", "that document")
+- **Intelligent Search**: Performs single or multiple searches depending on question complexity
+- **Source Citations**: Always includes sources with both titles and URIs
+- **Full Document Retrieval**: Can fetch complete documents on request
+- **Multiple Skills**: Exposes three distinct skills with appropriate artifacts:
+  - `document-qa`: Conversational question answering (default)
+  - `document-search`: Semantic search with structured results
+  - `document-retrieve`: Fetch complete documents by URI
+
+## Installation
+
+This package is not published to PyPI. Install it locally from the haiku.rag repository:
+
+```bash
+cd examples/a2a-server
+uv sync
+```
+
+This will install the package and all its dependencies, including `haiku.rag`.
+
+## Quick Start
+
+### Starting the A2A Server
+
+```bash
+# Start server with default database location (uses the same default as haiku-rag)
+uv run haiku-rag-a2a serve
+
+# Or specify a custom database path
+uv run haiku-rag-a2a serve --db /path/to/database
+
+# Start on custom host/port
+uv run haiku-rag-a2a serve --host 0.0.0.0 --port 8080
+```
+
+By default, the server uses the same database location as `haiku-rag`:
+- Linux: `~/.local/share/haiku.rag`
+- macOS: `~/Library/Application Support/haiku.rag`
+- Windows: `C:/Users/<USER>/AppData/Roaming/haiku.rag`
+
+### Interactive Client
+
+Test and interact with the A2A server using the built-in interactive client:
+
+```bash
+# Connect to local server
+uv run haiku-rag-a2a client
+
+# Connect to remote server
+uv run haiku-rag-a2a client --url https://example.com:8000
+```
+
+The interactive client provides:
+- Rich markdown rendering of agent responses
+- Conversation context across multiple turns
+- Agent card discovery and display
+- Compact artifact summaries
+
+## Python Usage
+
+```python
+from pathlib import Path
+from haiku_rag_a2a.a2a import create_a2a_app
+import uvicorn
+
+# Create A2A app
+app = create_a2a_app(Path("/path/to/database"))
+
+# Run with uvicorn
+uvicorn.run(app, host="127.0.0.1", port=8000)
+```
+
+## Security Examples
+
+The `security_examples/` directory contains examples for securing the A2A server:
+
+- `apikey_example.py` - Simple API key authentication
+- `oauth2_github.py` - GitHub Personal Access Token authentication
+- `oauth2_example.py` - Full OAuth2 with JWT verification
+
+## Architecture
+
+The A2A agent uses:
+
+- **FastA2A**: Python framework implementing the A2A protocol
+- **Pydantic AI**: Agent framework with tool support
+- **In-Memory Storage**: Context and message history storage (persists during server lifetime)
+- **Conversation State**: Full pydantic-ai message history serialized in A2A context
+
+## Configuration
+
+The server uses the same configuration as haiku.rag. You can specify a config file:
+
+```bash
+uv run haiku-rag-a2a serve --db /path/to/database --config haiku.rag.yaml
+```
+
+You can also control the maximum number of conversation contexts via the `--max-contexts` parameter (defaults to 1000).
+
+## Documentation
+
+See [a2a.md](./a2a.md) for detailed documentation including:
+- API examples
+- Security configuration
+- Docker deployment
+- Artifact specification
+
+## License
+
+MIT

examples/a2a-server/haiku_rag_a2a/a2a/__init__.py

@@ -9,7 +9,7 @@ from haiku.rag.config import AppConfig, Config
 from haiku.rag.graph_common import get_model
 
 from .context import load_message_history, save_message_history
-from .models import AgentDependencies, SearchResult
+from .models import A2AConfig, AgentDependencies, SearchResult
 from .prompts import A2A_SYSTEM_PROMPT
 from .skills import extract_question_from_task, get_agent_skills
 from .storage import LRUMemoryStorage
@@ -37,12 +37,14 @@ __all__ = [
     "extract_question_from_task",
     "get_agent_skills",
     "LRUMemoryStorage",
+    "A2AConfig",
 ]
 
 
 def create_a2a_app(
     db_path: Path,
     config: AppConfig = Config,
+    max_contexts: int = 1000,
     security_schemes: dict | None = None,
     security: list[dict[str, list[str]]] | None = None,
 ):
@@ -51,6 +53,7 @@ def create_a2a_app(
     Args:
         db_path: Path to the LanceDB database
         config: App configuration
+        max_contexts: Maximum number of conversations to keep in memory
         security_schemes: Optional security scheme definitions for the AgentCard
         security: Optional security requirements for the AgentCard
 
@@ -58,9 +61,7 @@ def create_a2a_app(
         A FastA2A ASGI application
     """
     base_storage = InMemoryStorage()
-    storage = LRUMemoryStorage(
+    storage = LRUMemoryStorage(storage=base_storage, max_contexts=max_contexts)
-        storage=base_storage, max_contexts=config.a2a.max_contexts
-    )
     broker = InMemoryBroker()
 
     # Create the agent with native search tool
@@ -123,7 +124,7 @@ def create_a2a_app(
     # Create FastA2A app with custom worker lifecycle
     @asynccontextmanager
     async def lifespan(app):
-        logger.info(f"Started A2A server (max contexts: {config.a2a.max_contexts})")
+        logger.info(f"Started A2A server (max contexts: {max_contexts})")
         async with app.task_manager:
             async with worker.run():
                 yield

examples/a2a-server/haiku_rag_a2a/a2a/models.py

@@ -3,6 +3,14 @@ from pydantic import BaseModel, Field
 from haiku.rag.client import HaikuRAG
 
 
+class A2AConfig(BaseModel):
+    """Configuration for A2A (Agent-to-Agent) protocol server."""
+
+    max_contexts: int = Field(
+        default=1000, description="Maximum number of conversations to keep in memory"
+    )
+
+
 class SearchResult(BaseModel):
     """Search result with both title and URI for A2A agent."""
 

examples/a2a-server/haiku_rag_a2a/a2a/worker.py

@@ -5,11 +5,11 @@ from pathlib import Path
 
 from pydantic_ai import Agent
 
-from haiku.rag.a2a.context import load_message_history, save_message_history
-from haiku.rag.a2a.models import AgentDependencies
-from haiku.rag.a2a.skills import extract_question_from_task
 from haiku.rag.client import HaikuRAG
 from haiku.rag.config import AppConfig, Config
+from haiku_rag_a2a.a2a.context import load_message_history, save_message_history
+from haiku_rag_a2a.a2a.models import AgentDependencies
+from haiku_rag_a2a.a2a.skills import extract_question_from_task
 
 try:
     from fasta2a import Worker  # type: ignore

examples/a2a-server/haiku_rag_a2a/cli.py

@@ -0,0 +1,85 @@
+import asyncio
+import logging
+from pathlib import Path
+
+import typer
+import uvicorn
+
+from haiku.rag.config import AppConfig, Config, load_yaml_config
+from haiku.rag.utils import get_default_data_dir
+from haiku_rag_a2a.a2a import create_a2a_app
+from haiku_rag_a2a.a2a.client import run_interactive_client
+
+logger = logging.getLogger(__name__)
+
+cli = typer.Typer(name="haiku-rag-a2a", no_args_is_help=True)
+
+
+@cli.command("serve", help="Start haiku.rag A2A (Agent-to-Agent) server")
+def serve(
+    db: Path | None = typer.Option(
+        None,
+        "--db",
+        help="Path to the database directory",
+    ),
+    config_file: Path | None = typer.Option(
+        None,
+        "--config",
+        help="Path to the configuration file",
+    ),
+    host: str = typer.Option(
+        "127.0.0.1",
+        "--host",
+        help="Host to bind A2A server to",
+    ),
+    port: int = typer.Option(
+        8000,
+        "--port",
+        help="Port to bind A2A server to",
+    ),
+    max_contexts: int = typer.Option(
+        1000,
+        "--max-contexts",
+        help="Maximum number of conversation contexts to keep in memory",
+    ),
+) -> None:
+    """Start the A2A server."""
+    config = Config
+    if config_file:
+        yaml_data = load_yaml_config(config_file)
+        config = AppConfig.model_validate(yaml_data)
+
+    if db is None:
+        db = get_default_data_dir()
+
+    if not db.exists():
+        typer.echo(f"Error: Database directory {db} does not exist")
+        raise typer.Exit(1)
+
+    logger.info(f"Starting A2A server on {host}:{port}")
+
+    app = create_a2a_app(db_path=db, config=config, max_contexts=max_contexts)
+    uvicorn_config = uvicorn.Config(
+        app,
+        host=host,
+        port=port,
+        log_level="info",
+    )
+    server = uvicorn.Server(uvicorn_config)
+    asyncio.run(server.serve())
+
+
+@cli.command("client", help="Run interactive client to chat with A2A server")
+def client(
+    url: str = typer.Option(
+        "http://localhost:8000",
+        "--url",
+        help="URL of the A2A server",
+    ),
+):
+    """Run the interactive A2A client."""
+    asyncio.run(run_interactive_client(url))
+
+
+if __name__ == "__main__":
+    cli()

examples/a2a-server/pyproject.toml

@@ -0,0 +1,27 @@
+[project]
+name = "haiku-rag-a2a"
+description = "A2A protocol server for haiku.rag - Conversational agent interface"
+version = "0.1.0"
+authors = [{ name = "Yiorgis Gozadinos", email = "ggozadinos@gmail.com" }]
+license = { text = "MIT" }
+readme = "README.md"
+requires-python = ">=3.12"
+keywords = ["RAG", "a2a", "agent", "conversational-ai"]
+
+dependencies = [
+    "haiku.rag>=0.14.0",
+    "fasta2a>=0.1.0",
+    "pydantic-ai-slim[a2a]>=1.11.1",
+    "rich>=14.2.0",
+    "httpx>=0.28.1",
+]
+
+[project.scripts]
+haiku-rag-a2a = "haiku_rag_a2a.cli:cli"
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["haiku_rag_a2a"]

examples/a2a-server/security_examples/apikey_example.py

@@ -22,12 +22,11 @@ Usage:
 import os
 from pathlib import Path
 
+from haiku_rag_a2a.a2a import create_a2a_app
 from starlette.exceptions import HTTPException
 from starlette.responses import JSONResponse
 from starlette.status import HTTP_401_UNAUTHORIZED
 
-from haiku.rag.a2a import create_a2a_app
-
 # API Key Configuration - In production, use environment variables or a secure key store
 API_KEY_NAME = "X-API-Key"
 VALID_API_KEY = os.getenv("API_KEY", "demo-key-12345")

examples/a2a-server/security_examples/oauth2_example.py

@@ -35,6 +35,7 @@ Usage:
 import os
 from pathlib import Path
 
+from haiku_rag_a2a.a2a import create_a2a_app
 from jose import JWTError, jwt
 from starlette.exceptions import HTTPException
 from starlette.responses import JSONResponse
@@ -44,8 +45,6 @@ from starlette.status import (
     HTTP_500_INTERNAL_SERVER_ERROR,
 )
 
-from haiku.rag.a2a import create_a2a_app
-
 # OAuth2 Configuration
 OAUTH2_TOKEN_URL = os.getenv(
     "OAUTH2_TOKEN_URL", "https://your-auth.example.com/oauth/token"

examples/a2a-server/security_examples/oauth2_github.py

@@ -24,12 +24,11 @@ import os
 from pathlib import Path
 
 import httpx
+from haiku_rag_a2a.a2a import create_a2a_app
 from starlette.exceptions import HTTPException
 from starlette.responses import JSONResponse
 from starlette.status import HTTP_401_UNAUTHORIZED
 
-from haiku.rag.a2a import create_a2a_app
-
 # Configuration
 GITHUB_API_URL = "https://api.github.com"
 ALLOWED_TOKENS = (

examples/a2a-server/tests/test_a2a.py

@@ -5,6 +5,13 @@ import pytest
 pytest.importorskip("fasta2a")
 
 from fasta2a.schema import Message, TextPart  # noqa: E402
+from haiku_rag_a2a.a2a import (
+    extract_question_from_task,
+    get_agent_skills,
+    load_message_history,
+    save_message_history,
+)
+from haiku_rag_a2a.a2a.storage import LRUMemoryStorage
 from pydantic_ai.messages import (  # noqa: E402
     ModelMessage,
     ModelRequest,
@@ -16,13 +23,6 @@ from pydantic_ai.messages import (
     TextPart as AITextPart,
 )
 
-from haiku.rag.a2a import (
-    extract_question_from_task,
-    get_agent_skills,
-    load_message_history,
-    save_message_history,
-)
-from haiku.rag.a2a.storage import LRUMemoryStorage
 from haiku.rag.client import HaikuRAG
 
 
@@ -209,7 +209,7 @@ async def test_lru_memory_storage_access_order():
 @pytest.mark.asyncio
 async def test_a2a_app_creation(temp_db_path):
     """Test that A2A app can be created successfully."""
-    from haiku.rag.a2a import create_a2a_app
+    from haiku_rag_a2a.a2a import create_a2a_app
 
     # Create a test database
     async with HaikuRAG(temp_db_path) as client:
@@ -230,7 +230,7 @@ async def test_a2a_app_creation(temp_db_path):
 @pytest.mark.asyncio
 async def test_a2a_app_has_skills(temp_db_path):
     """Test that A2A app exposes skills describing its capabilities."""
-    from haiku.rag.a2a import create_a2a_app
+    from haiku_rag_a2a.a2a import create_a2a_app
 
     # Create a test database
     async with HaikuRAG(temp_db_path) as client:
@@ -291,6 +291,7 @@ def test_get_agent_skills():
 @pytest.mark.asyncio
 async def test_build_artifacts_for_search():
     """Test that search operations produce structured search artifacts."""
+    from haiku_rag_a2a.a2a.worker import ConversationalWorker
     from pydantic_ai.messages import (
         ModelRequest,
         ModelResponse,
@@ -298,8 +299,6 @@ async def test_build_artifacts_for_search():
         ToolReturnPart,
     )
 
-    from haiku.rag.a2a.worker import ConversationalWorker
-
     class MockResult:
         output = "Found 1 relevant results:\n\n1. *Score: 0.9* | **test**\nresult"
 
@@ -350,6 +349,7 @@ async def test_build_artifacts_for_search():
 @pytest.mark.asyncio
 async def test_build_artifacts_for_retrieve():
     """Test that retrieve operations produce document artifacts."""
+    from haiku_rag_a2a.a2a.worker import ConversationalWorker
     from pydantic_ai.messages import (
         ModelRequest,
         ModelResponse,
@@ -357,8 +357,6 @@ async def test_build_artifacts_for_retrieve():
         ToolReturnPart,
     )
 
-    from haiku.rag.a2a.worker import ConversationalWorker
-
     class MockResult:
         output = "Document content"
 
@@ -407,6 +405,7 @@ async def test_build_artifacts_for_retrieve():
 @pytest.mark.asyncio
 async def test_build_artifacts_for_multiple_searches():
     """Test that multiple searches each get their own artifact with correct results."""
+    from haiku_rag_a2a.a2a.worker import ConversationalWorker
     from pydantic_ai.messages import (
         ModelRequest,
         ModelResponse,
@@ -415,8 +414,6 @@ async def test_build_artifacts_for_multiple_searches():
     )
     from pydantic_ai.messages import TextPart as AITextPart
 
-    from haiku.rag.a2a.worker import ConversationalWorker
-
     class MockResult:
         output = "Answer based on multiple searches"
 
@@ -513,11 +510,10 @@ async def test_build_artifacts_for_multiple_searches():
 @pytest.mark.asyncio
 async def test_qa_artifact_for_conversational_messages():
     """Test that conversational Q&A messages always create qa_result artifacts."""
+    from haiku_rag_a2a.a2a.worker import ConversationalWorker
     from pydantic_ai.messages import ModelResponse
     from pydantic_ai.messages import TextPart as AITextPart
 
-    from haiku.rag.a2a.worker import ConversationalWorker
-
     class MockResult:
         output = "Hello! How can I help you?"
 
@@ -552,6 +548,7 @@ async def test_qa_artifact_for_conversational_messages():
 @pytest.mark.asyncio
 async def test_build_artifacts_for_qa():
     """Test that Q&A operations produce artifacts for each tool call."""
+    from haiku_rag_a2a.a2a.worker import ConversationalWorker
     from pydantic_ai.messages import (
         ModelRequest,
         ModelResponse,
@@ -560,8 +557,6 @@ async def test_build_artifacts_for_qa():
     )
     from pydantic_ai.messages import TextPart as AITextPart
 
-    from haiku.rag.a2a.worker import ConversationalWorker
-
     class MockResult:
         output = "This is the answer"
 

examples/docker/README.md

@@ -1,6 +1,6 @@
 # haiku.rag Docker Compose Example
 
-Run haiku.rag with file monitoring, MCP server, and A2A agent.
+Run haiku.rag with file monitoring and MCP server.
 
 ## Quick Start
 
@@ -23,14 +23,10 @@ docker compose exec haiku-rag haiku-rag search "your query"
 
 # Ask questions
 docker compose exec haiku-rag haiku-rag ask "What is haiku.rag?"
-
-# A2A interactive client
-docker compose exec haiku-rag haiku-rag a2aclient --url http://localhost:8000
 ```
 
 ## Ports
 
-- `8000` - A2A agent
 - `8001` - MCP server
 
 ## Configuration
@@ -52,4 +48,3 @@ docker compose up -d
 - [Configuration](https://ggozad.github.io/haiku.rag/configuration/)
 - [CLI Commands](https://ggozad.github.io/haiku.rag/cli/)
 - [MCP Server](https://ggozad.github.io/haiku.rag/mcp/)
-- [A2A Agent](https://ggozad.github.io/haiku.rag/a2a/)

examples/docker/docker-compose.yml

@@ -5,7 +5,6 @@ services:
       dockerfile: docker/Dockerfile
     container_name: haiku-rag
     ports:
-      - "8000:8000"  # A2A server
       - "8001:8001"  # MCP server
     volumes:
       - ./data:/data  # Persist database
@@ -19,9 +18,3 @@ services:
       - CO_API_KEY=${CO_API_KEY}
 
     restart: unless-stopped
-    healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
-      interval: 30s
-      timeout: 10s
-      retries: 3
-      start_period: 40s

haiku_rag_slim/README.md

@@ -2,7 +2,7 @@
 
 Retrieval-Augmented Generation (RAG) library built on LanceDB - Core package with minimal dependencies.
 
-`haiku.rag-slim` is the core package for users who want to install only the dependencies they need. Document processing (docling), rerankers, and A2A support are all optional extras.
+`haiku.rag-slim` is the core package for users who want to install only the dependencies they need. Document processing (docling), and reranker support are all optional extras.
 
 **For most users, we recommend installing [`haiku.rag`](https://pypi.org/project/haiku.rag/) instead**, which includes all features out of the box.
 
@@ -48,8 +48,6 @@ Adds support for 40+ file formats including PDF, DOCX, HTML, and more.
 - `bedrock` - AWS Bedrock
 - `vertexai` - Google Vertex AI
 
-**Agent Protocol:**
-- `a2a` - Agent-to-Agent protocol
 
 ```bash
 # Common combinations
@@ -64,7 +62,6 @@ See the main [`haiku.rag`](https://github.com/ggozad/haiku.rag) repository for:
 - CLI examples
 - Python API usage
 - MCP server setup
-- A2A agent configuration
 
 ## Documentation
 

haiku_rag_slim/haiku/rag/app.py

@@ -455,9 +455,6 @@ class HaikuRAGApp:
         enable_mcp: bool = True,
         mcp_transport: str | None = None,
         mcp_port: int = 8001,
-        enable_a2a: bool = False,
-        a2a_host: str = "127.0.0.1",
-        a2a_port: int = 8000,
     ):
         """Start the server with selected services."""
         async with HaikuRAG(self.db_path, config=self.config) as client:
@@ -485,33 +482,6 @@ class HaikuRAGApp:
                 mcp_task = asyncio.create_task(run_mcp())
                 tasks.append(mcp_task)
 
-            # Start A2A server if enabled
-            if enable_a2a:
-                try:
-                    from haiku.rag.a2a import create_a2a_app
-                except ImportError as e:
-                    logger.error(f"Failed to import A2A: {e}")
-                    return
-
-                import uvicorn
-
-                logger.info(f"Starting A2A server on {a2a_host}:{a2a_port}")
-
-                async def run_a2a():
-                    app = create_a2a_app(db_path=self.db_path, config=self.config)
-                    uvicorn_config = uvicorn.Config(
-                        app,
-                        host=a2a_host,
-                        port=a2a_port,
-                        log_level="warning",
-                        access_log=False,
-                    )
-                    server = uvicorn.Server(uvicorn_config)
-                    await server.serve()
-
-                a2a_task = asyncio.create_task(run_a2a())
-                tasks.append(a2a_task)
-
             if not tasks:
                 logger.warning("No services enabled")
                 return

haiku_rag_slim/haiku/rag/cli.py

@@ -401,7 +401,7 @@ def download_models_cmd():
 
 @cli.command(
     "serve",
-    help="Start haiku.rag server. Use --monitor, --mcp, and/or --a2a to enable services.",
+    help="Start haiku.rag server. Use --monitor and/or --mcp to enable services.",
 )
 def serve(
     db: Path | None = typer.Option(
@@ -429,27 +429,12 @@ def serve(
         "--mcp-port",
         help="Port to bind MCP server to (ignored with --stdio)",
     ),
-    a2a: bool = typer.Option(
-        False,
-        "--a2a",
-        help="Enable A2A (Agent-to-Agent) server",
-    ),
-    a2a_host: str = typer.Option(
-        "127.0.0.1",
-        "--a2a-host",
-        help="Host to bind A2A server to",
-    ),
-    a2a_port: int = typer.Option(
-        8000,
-        "--a2a-port",
-        help="Port to bind A2A server to",
-    ),
 ) -> None:
     """Start the server with selected services."""
     # Require at least one service flag
-    if not (monitor or mcp or a2a):
+    if not (monitor or mcp):
         typer.echo(
-            "Error: At least one service flag (--monitor, --mcp, or --a2a) must be specified"
+            "Error: At least one service flag (--monitor or --mcp) must be specified"
         )
         raise typer.Exit(1)
 
@@ -467,34 +452,9 @@ def serve(
             enable_mcp=mcp,
             mcp_transport=transport,
             mcp_port=mcp_port,
-            enable_a2a=a2a,
-            a2a_host=a2a_host,
-            a2a_port=a2a_port,
         )
     )
 
 
-@cli.command(
-    "a2aclient", help="Run interactive client to chat with haiku.rag's A2A server"
-)
-def a2aclient(
-    url: str = typer.Option(
-        "http://localhost:8000",
-        "--url",
-        help="Base URL of the A2A server",
-    ),
-):
-    try:
-        from haiku.rag.a2a.client import run_interactive_client
-    except ImportError:
-        typer.echo(
-            "Error: A2A support requires the 'a2a' extra. "
-            "Install with: uv pip install 'haiku.rag[a2a]'"
-        )
-        raise typer.Exit(1)
-
-    asyncio.run(run_interactive_client(url=url))
-
-
 if __name__ == "__main__":
     cli()

haiku_rag_slim/haiku/rag/config/__init__.py

@@ -6,7 +6,6 @@ from haiku.rag.config.loader import (
     load_yaml_config,
 )
 from haiku.rag.config.models import (
-    A2AConfig,
     AppConfig,
     EmbeddingsConfig,
     LanceDBConfig,
@@ -35,7 +34,6 @@ __all__ = [
     "OllamaConfig",
     "VLLMConfig",
     "ProvidersConfig",
-    "A2AConfig",
     "find_config_file",
     "load_yaml_config",
     "generate_default_config",

haiku_rag_slim/haiku/rag/config/loader.py

@@ -84,5 +84,4 @@ def generate_default_config() -> dict:
                 "research_base_url": "",
             },
         },
-        "a2a": {"max_contexts": 1000},
     }

haiku_rag_slim/haiku/rag/config/models.py

@@ -76,10 +76,6 @@ class ProvidersConfig(BaseModel):
     vllm: VLLMConfig = Field(default_factory=VLLMConfig)
 
 
-class A2AConfig(BaseModel):
-    max_contexts: int = 1000
-
-
 class AppConfig(BaseModel):
     environment: str = "production"
     storage: StorageConfig = Field(default_factory=StorageConfig)
@@ -91,4 +87,3 @@ class AppConfig(BaseModel):
     research: ResearchConfig = Field(default_factory=ResearchConfig)
     processing: ProcessingConfig = Field(default_factory=ProcessingConfig)
     providers: ProvidersConfig = Field(default_factory=ProvidersConfig)
-    a2a: A2AConfig = Field(default_factory=A2AConfig)

haiku_rag_slim/pyproject.toml

@@ -45,8 +45,6 @@ voyageai = ["voyageai>=0.3.5"]
 mxbai = ["mxbai-rerank>=0.1.6"]
 cohere = ["cohere>=5.0.0"]
 zeroentropy = ["zeroentropy>=0.1.0a6"]
-# Agent protocols
-a2a = ["fasta2a>=0.1.0", "pydantic-ai-slim[a2a]"]
 # Model providers (delegated to pydantic-ai-slim)
 anthropic = ["pydantic-ai-slim[anthropic]"]
 groq = ["pydantic-ai-slim[groq]"]

mkdocs.yml

@@ -65,7 +65,6 @@ nav:
       - Agents: agents.md
       - Server: server.md
       - MCP: mcp.md
-      - A2A: a2a.md
       - Benchmarks: benchmarks.md
 markdown_extensions:
   - admonition

pyproject.toml

@@ -21,7 +21,7 @@ classifiers = [
     "Typing :: Typed",
 ]
 
-dependencies = ["haiku.rag-slim[docling,voyageai,mxbai,a2a,cohere,zeroentropy]"]
+dependencies = ["haiku.rag-slim[docling,voyageai,mxbai,cohere,zeroentropy]"]
 
 [project.scripts]
 haiku-rag = "haiku.rag.cli:cli"
@@ -100,6 +100,8 @@ pythonVersion = "3.12"
 [tool.pytest.ini_options]
 asyncio_default_fixture_loop_scope = "session"
 asyncio_mode = "auto"
+testpaths = ["tests"]
+norecursedirs = ["examples", "docs", "evaluations", ".git", ".venv"]
 
 # pyproject.toml
 filterwarnings = ["error", "ignore::UserWarning", "ignore::DeprecationWarning"]

tests/test_app.py

@@ -211,7 +211,6 @@ async def test_serve_mcp_only(app: HaikuRAGApp, monkeypatch, transport):
                 enable_monitor=False,
                 enable_mcp=True,
                 mcp_transport=transport,
-                enable_a2a=False,
             )
         except asyncio.CancelledError:
             pass
@@ -245,50 +244,16 @@ async def test_serve_monitor_only(app: HaikuRAGApp, monkeypatch):
 
     with patch("haiku.rag.app.HaikuRAG", return_value=mock_client):
         try:
-            await app.serve(enable_monitor=True, enable_mcp=False, enable_a2a=False)
+            await app.serve(enable_monitor=True, enable_mcp=False)
         except asyncio.CancelledError:
             pass
 
     assert len(created_tasks) == 1
 
 
-@pytest.mark.asyncio
-async def test_serve_a2a_only(app: HaikuRAGApp, monkeypatch):
-    """Test the serve method with A2A server only."""
-    pytest.importorskip("fasta2a")
-    created_tasks = []
-    original_create_task = asyncio.create_task
-
-    def track_task(coro):
-        task = original_create_task(coro)
-        created_tasks.append(task)
-        task.cancel()
-        return task
-
-    monkeypatch.setattr("haiku.rag.app.asyncio.create_task", track_task)
-    monkeypatch.setattr(
-        "haiku.rag.app.asyncio.gather", AsyncMock(side_effect=asyncio.CancelledError)
-    )
-
-    mock_client = AsyncMock()
-    mock_client.__aenter__.return_value = mock_client
-
-    mock_a2a_app = MagicMock()
-
-    with patch("haiku.rag.app.HaikuRAG", return_value=mock_client):
-        with patch("haiku.rag.a2a.create_a2a_app", return_value=mock_a2a_app):
-            try:
-                await app.serve(enable_monitor=False, enable_mcp=False, enable_a2a=True)
-            except asyncio.CancelledError:
-                pass
-
-    assert len(created_tasks) == 1
-
-
 @pytest.mark.asyncio
 async def test_serve_all_services(app: HaikuRAGApp, monkeypatch):
     """Test the serve method with all services enabled."""
-    pytest.importorskip("fasta2a")
     created_tasks = []
     original_create_task = asyncio.create_task
 
@@ -300,7 +265,6 @@ async def test_serve_all_services(app: HaikuRAGApp, monkeypatch):
 
     mock_server = AsyncMock()
     mock_watcher = AsyncMock()
-    mock_a2a_app = MagicMock()
 
     monkeypatch.setattr(
         "haiku.rag.app.create_mcp_server", MagicMock(return_value=mock_server)
@@ -317,13 +281,12 @@ async def test_serve_all_services(app: HaikuRAGApp, monkeypatch):
     mock_client.__aenter__.return_value = mock_client
 
     with patch("haiku.rag.app.HaikuRAG", return_value=mock_client):
-        with patch("haiku.rag.a2a.create_a2a_app", return_value=mock_a2a_app):
+        try:
-            try:
+            await app.serve(enable_monitor=True, enable_mcp=True)
-                await app.serve(enable_monitor=True, enable_mcp=True, enable_a2a=True)
+        except asyncio.CancelledError:
-            except asyncio.CancelledError:
+            pass
-                pass
 
-    assert len(created_tasks) == 3
+    assert len(created_tasks) == 2
 
 
 @pytest.mark.asyncio

tests/test_cli.py

@@ -196,7 +196,6 @@ def test_serve_mcp_only():
         _, kwargs = mock_app_instance.serve.call_args
         assert kwargs["enable_monitor"] is False
         assert kwargs["enable_mcp"] is True
-        assert kwargs["enable_a2a"] is False
         assert kwargs["mcp_transport"] is None
         assert kwargs["mcp_port"] == 8001
 
@@ -230,26 +229,6 @@ def test_serve_monitor_only():
         _, kwargs = mock_app_instance.serve.call_args
         assert kwargs["enable_monitor"] is True
         assert kwargs["enable_mcp"] is False
-        assert kwargs["enable_a2a"] is False
-
-
-def test_serve_a2a_only():
-    """Test serve command with A2A only."""
-    with patch("haiku.rag.cli.HaikuRAGApp") as mock_app:
-        mock_app_instance = MagicMock()
-        mock_app_instance.serve = AsyncMock()
-        mock_app.return_value = mock_app_instance
-
-        result = runner.invoke(cli, ["serve", "--a2a"])
-
-        assert result.exit_code == 0
-        mock_app_instance.serve.assert_called_once()
-        _, kwargs = mock_app_instance.serve.call_args
-        assert kwargs["enable_monitor"] is False
-        assert kwargs["enable_mcp"] is False
-        assert kwargs["enable_a2a"] is True
-        assert kwargs["a2a_host"] == "127.0.0.1"
-        assert kwargs["a2a_port"] == 8000
 
 
 def test_serve_all_services():
@@ -259,32 +238,28 @@ def test_serve_all_services():
         mock_app_instance.serve = AsyncMock()
         mock_app.return_value = mock_app_instance
 
-        result = runner.invoke(cli, ["serve", "--monitor", "--mcp", "--a2a"])
+        result = runner.invoke(cli, ["serve", "--monitor", "--mcp"])
 
         assert result.exit_code == 0
         mock_app_instance.serve.assert_called_once()
         _, kwargs = mock_app_instance.serve.call_args
         assert kwargs["enable_monitor"] is True
         assert kwargs["enable_mcp"] is True
-        assert kwargs["enable_a2a"] is True
 
 
 def test_serve_custom_ports():
-    """Test serve command with custom ports."""
+    """Test serve command with custom MCP port."""
     with patch("haiku.rag.cli.HaikuRAGApp") as mock_app:
         mock_app_instance = MagicMock()
         mock_app_instance.serve = AsyncMock()
         mock_app.return_value = mock_app_instance
 
-        result = runner.invoke(
+        result = runner.invoke(cli, ["serve", "--mcp", "--mcp-port", "9000"])
-            cli, ["serve", "--mcp", "--mcp-port", "9000", "--a2a", "--a2a-port", "9001"]
-        )
 
         assert result.exit_code == 0
         mock_app_instance.serve.assert_called_once()
         _, kwargs = mock_app_instance.serve.call_args
         assert kwargs["mcp_port"] == 9000
-        assert kwargs["a2a_port"] == 9001
 
 
 def test_serve_stdio_without_mcp():