3 ms·

Atomic Agents 0.1.44 Released

DeadlyPretzel 2y ago
For those who don't know yet, Atomic Agents ( https://github.com/KennyVaneetvelde/atomic_agents https://github.com/KennyVaneetvelde/atomic_agents ) is designed to be modular, extensible, and easy to use. Components in the Atomic Agents Framework should always be as small and single-purpose as possible, similar to design system components in Atomic Design. Even though Atomic Design cannot be directly applied to AI agent architecture, a lot of ideas were taken from it. The resulting framework provides a set of tools and agents that can be combined to create powerful applications. The framework is built on top of Instructor and uses Pydantic for data validation and serialization. For those who have been following it for a bit, it just got a lot easier to build new agents using any client supported by Instructor, including local agents. I highly recommend checking out: - The basic custom chatbot example: https://github.com/KennyVaneetvelde/atomic_agents/blob/main/examples/notebooks/quickstart.ipynb https://github.com/KennyVaneetvelde/atomic_agents/blob/main/... - Yelp agent to help find restaurants on yelp: https://github.com/KennyVaneetvelde/atomic_agents/blob/main/examples/notebooks/yelp_agent.ipynb https://github.com/KennyVaneetvelde/atomic_agents/blob/main/... This demo essentially shows how an agent in Atomic Agents can be given a schema and figure out the best way on its own to ask the user the right questions in order to gather the necessary information for performing the API call. This logic can essentially be applied to any filterable API or endpoint, ... such as for a webshop's products (hint hint, product idea) - Deep multi-agent research example (like perplexity): https://github.com/KennyVaneetvelde/atomic_agents/tree/main/examples/deep_research_multi_agent https://github.com/KennyVaneetvelde/atomic_agents/tree/main/... - Agent orchestration demo (in other words, letting an agent outsource tasks to other agents): https://github.com/KennyVaneetvelde/atomic_agents/blob/main/examples/notebooks/multi_agent_quickstart.ipynb https://github.com/KennyVaneetvelde/atomic_agents/blob/main/... - Easily sharing dynamic context between two atomic agents: https://github.com/KennyVaneetvelde/atomic_agents/blob/main/examples/shared_context.py https://github.com/KennyVaneetvelde/atomic_agents/blob/main/... More examples: https://github.com/KennyVaneetvelde/atomic_agents/tree/main/examples https://github.com/KennyVaneetvelde/atomic_agents/tree/main/... Docs: https://github.com/KennyVaneetvelde/atomic_agents/tree/main/docs https://github.com/KennyVaneetvelde/atomic_agents/tree/main/...

For those who don't know yet, Atomic Agents ( https://github.com/KennyVaneetvelde/atomic_agents https://github.com/KennyVaneetvelde/atomic_agents ) is designed to be modular, extensible, and easy to use. Components in the Atomic Agents Framework should always be as small and single-purpose as possible, similar to design system components in Atomic Design. Even though Atomic Design cannot be directly applied to AI agent architecture, a lot of ideas were taken from it. The resulting framework provides a set of tools and agents that can be combined to create powerful applications. The framework is built on top of Instructor and uses Pydantic for data validation and serialization. For those who have been following it for a bit, it just got a lot easier to build new agents using any client supported by Instructor, including local agents. I highly recommend checking out: - The basic custom chatbot example: https://github.com/KennyVaneetvelde/atomic_agents/blob/main/examples/notebooks/quickstart.ipynb https://github.com/KennyVaneetvelde/atomic_agents/blob/main/... - Yelp agent to help find restaurants on yelp: https://github.com/KennyVaneetvelde/atomic_agents/blob/main/examples/notebooks/yelp_agent.ipynb https://github.com/KennyVaneetvelde/atomic_agents/blob/main/... This demo essentially shows how an agent in Atomic Agents can be given a schema and figure out the best way on its own to ask the user the right questions in order to gather the necessary information for performing the API call. This logic can essentially be applied to any filterable API or endpoint, ... such as for a webshop's products (hint hint, product idea) - Deep multi-agent research example (like perplexity): https://github.com/KennyVaneetvelde/atomic_agents/tree/main/examples/deep_research_multi_agent https://github.com/KennyVaneetvelde/atomic_agents/tree/main/... - Agent orchestration demo (in other words, letting an agent outsource tasks to other agents): https://github.com/KennyVaneetvelde/atomic_agents/blob/main/examples/notebooks/multi_agent_quickstart.ipynb https://github.com/KennyVaneetvelde/atomic_agents/blob/main/... - Easily sharing dynamic context between two atomic agents: https://github.com/KennyVaneetvelde/atomic_agents/blob/main/examples/shared_context.py https://github.com/KennyVaneetvelde/atomic_agents/blob/main/... More examples: https://github.com/KennyVaneetvelde/atomic_agents/tree/main/examples https://github.com/KennyVaneetvelde/atomic_agents/tree/main/... Docs: https://github.com/KennyVaneetvelde/atomic_agents/tree/main/docs https://github.com/KennyVaneetvelde/atomic_agents/tree/main/...