Connic
Back to BlogProduct Release

Composer SDK: Better Agent Development Tooling

No more manual uploads or YAML guessing. The Composer SDK brings scaffolding, validation, hot-reload testing, and CLI deployments.

December 27, 20255 min read

Connic has always let you define agents in YAML and write tools in Python. But until now, getting everything set up and deployed meant a lot of manual work: uploading files by hand, setting up project structures yourself, and hoping you didn't have a typo in your YAML. Today we're releasing the Composer SDK to fix all of that.

What Was the Problem?

Building agents on Connic worked, but the developer experience had friction:

  • Manual uploads: Every time you changed an agent or tool, you had to upload the files yourself
  • No scaffolding: You had to create the folder structure and boilerplate files from scratch
  • No local validation: You'd only find out about YAML syntax errors or missing tools after uploading
  • Slow iteration: The upload-test-fix cycle was tedious, especially for small changes

The SDK solves each of these with proper CLI tooling.

What the SDK Does

Install it from PyPI:

Terminal
pip install connic-composer-sdk

Then you get a set of commands that streamline the entire workflow:

connic init - Project Scaffolding

No more setting up folders manually. One command creates a complete project structure with example agents, tools, middleware, and a README explaining everything.

Terminal
$ connic init my-agents

✓ Initialized Connic project in /Users/you/my-agents

Created files:
  agents/assistant.yaml          (LLM agent)
  agents/invoice-processor.yaml  (LLM agent with retry)
  agents/tax-calculator.yaml     (Tool agent)
  agents/document-pipeline.yaml  (Sequential agent)
  agents/orchestrator.yaml       (Orchestrator with trigger_agent)
  agents/knowledge-agent.yaml    (Knowledge agent with RAG)
  tools/calculator.py
  middleware/assistant.py
  .gitignore
  requirements.txt
  README.md

The generated examples cover all agent types (LLM, Tool, Sequential) plus patterns like orchestration and knowledge retrieval. Delete what you don't need, modify the rest.

connic dev - Local Validation

Catch errors before uploading. The SDK loads all your agents and tools locally, validates the YAML, checks that referenced tools exist, and shows you exactly what would be deployed.

Terminal
$ connic dev

Connic Composer SDK - Development Mode

Discovering tools...
  Found 3 tools in 1 modules:
    calculator: add, multiply, calculate_tax

Discovering middlewares...
  Found middlewares for 1 agents:
    assistant: before, after

Loading agents...
  Found 3 agents:

  ┌─ assistant [LLM]
  │  Description: A helpful assistant with calculator tools
  │  Model: gemini/gemini-2.5-flash
  │  Tools: calculator.add, calculator.multiply, calculator.calculate_tax
  └─

  ┌─ tax-calculator [Tool]
  │  Description: Calculates tax directly
  │  Tool: calculator.calculate_tax
  └─

  ┌─ document-pipeline [Sequential]
  │  Description: Processes documents through extraction and validation
  │  Chain: assistant → invoice-processor
  └─

✓ Project validation complete

If something's wrong (a misspelled tool reference, invalid YAML, missing middleware) you'll see it immediately instead of after uploading.

connic test - Hot Reload Testing

This is the big one. Start a test session and your local files sync to a cloud environment automatically. Make a change, save the file, and it's live in 2-5 seconds. No manual upload needed.

Terminal
$ connic test

Watching for file changes...
  Press Ctrl+C to stop

────────────────────────────────────────────────────────────
  View and trigger your agents: https://connic.co/projects/...
────────────────────────────────────────────────────────────

[14:32:15] Detected change: assistant.yaml
[14:32:16] Files changed, uploading...
[14:32:17] Uploaded 2,847 bytes

The CLI watches your agents/, tools/, middleware/, and requirements.txt. Test using connectors in the dashboard, iterate on your code, and see changes reflected almost instantly.

connic deploy - CLI Deployment

Deploy directly from your terminal or CI/CD pipeline, without going through the dashboard.

Terminal
connic login              # Save credentials
connic deploy             # Deploy to default environment
connic deploy --env <id>  # Deploy to specific environment

Combine this with Git integration and pushing to your repo triggers automatic deployments. No CLI needed at all for the normal workflow.

New: Import Predefined Tools

Along with the SDK, we've added the ability to import predefined tools into your own custom tools. This means you can build on top of capabilities like knowledge retrieval, web search, or agent orchestration in your Python code.

tools/my_tools.py
from connic.predefined import query_knowledge, store_knowledge

async def lookup_and_summarize(query: str) -> str:
    """Search the knowledge base and return a summary."""
    # Use the predefined knowledge tool
    results = await query_knowledge(
        query=query,
        namespace="documents",
        top_k=5
    )
    
    # Process and return the results
    if not results:
        return "No relevant information found."
    
    return "\n".join([r["content"] for r in results])

Previously, predefined tools could only be used directly by agents. Now you can compose them into more complex tool logic: query knowledge, process the results, make decisions, and return exactly what you need.

Get Started

The SDK is available now. If you've been using Connic without it, the transition is simple: your existing YAML configs and Python tools work exactly the same. You just get better tooling around them.

Terminal
pip install connic-composer-sdk
connic init my-agents  # Or use with your existing project
cd my-agents
connic dev             # Validate everything locally
connic test            # Start hot-reload testing

Check out the quickstart guide for a complete walkthrough, or see the SDK documentation for the full reference.