Metadata-Version: 2.2
Name: agentstr
Version: 0.0.8
Summary: A library for collaborative AI agents
Home-page: https://github.com/synvya/agentstr
Author: Alejandro Gil
Author-email: Alejandro Gil <info@synvya.com>
License: MIT
Project-URL: Homepage, https://github.com/synvya/agentstr
Project-URL: Documentation, https://github.com/synvya/agentstr#readme
Project-URL: BugTracker, https://github.com/synvya/agentstr/issues
Keywords: AI,agents,collaboration,library
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.6
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: phi>=2.7.0
Requires-Dist: openai>=1.50.0
Dynamic: author
Dynamic: home-page
Dynamic: requires-python

AgentStr
========
AgentStr is an extension of [Phidata](https://www.phidata.com) AI agents that allows for agents to communicate with other agents in separate computers using the Nostr communication protocol.

The goal is for Agent A operated by Company A to be able to work with Agent B operated by Company B to achieve a common goal. For example: Company A wants to buy a product sold by Company B so Agent A and Agent B can coordinate and execute the transaction. 

# License
This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details.

# Current status
The library is in its infancy.

Done:
- Workflow to package and distribute the library
- Create AgentStr as an extension of Phidata Agent
- Test the addition of new properties and capabilities for the AgentStr class

To be done:
- Incorporate Nostr capabilities
- Create unique public / private key identifiers for agent instances
- Send and retreive messages via Nostr
- Expose capabilities via Nostr
- Agent B retrieves capabilities exposed by Agent A
- Agent B coordinates transaction with Agent A

# Installation
AgentStr is offered as a python library available at https://pypi.org/project/agentstr/. 

Here is an example on how to use the library:

1. Create a new python environment for your app
    ```
    cd ~/
    python3 -m venv mysampleapp
    source mysampleapp/bin/acticate
    ```
2. Install the agentstr library
    ```
    cd ~/mysampleapp
    pip install agentstr
    ```
3. Create a new python file
    ```
    touch main.py
    ```
4. Copy paste this code to the main.py file
    ```
    from agentstr.core import AgentStr
    # Create the agent
    agent = AgentStr("Synvya Inc", "Seller")

    # Test AgentStr new capabilities
    print(f"Public key: {agent.get_public_key()}\nPrivate key: {agent.get_private_key()}")
    print(f"Company: {agent.get_company()}\nRole: {agent.get_role()}")

    # Test phidata inherited capabilities
    agent.print_response("Write two sentence poem for  the love between the sun and the moon.") 
    ```
5. Run the code
    ```
    ~/mysampleapp/bin/python main.py
    ```

# Contributing
Refer to [CONTRIBUTING.md](CONTRIBUTING.md) for specific instructions on installation instructions for developers and how to contribute.

# Acknowledgments
- [Phidata](https://www.phidata.com/.com/) - For building robust AI agents.
