GPT models + function calling — AGNT's llm_gateway supports OpenAI as a provider.
OpenAI API + AGNT
GPT models + function calling — AGNT's llm_gateway supports OpenAI as a provider.
Official sourceWhat it is
The OpenAI API provides access to GPT models with function calling, structured outputs, and streaming. It is the most widely adopted LLM API and the de facto standard for tool-calling schema definitions.
For AGNT composition: OpenAI's embedding API (`text-embedding-3-small`) powers the semantic recall engine, and the function-calling format is the baseline schema AGNT's tool executor normalizes against.
Where AGNT fits
- AGNT's semantic recall engine uses OpenAI's embedding API to embed user messages and search against stored memory via pgvector cosine distance. Every non-trivial message is embedded before recall runs.
- The `llm_gateway.py` supports OpenAI as an alternative provider. Builders who prefer GPT for specific tasks can route through the same gateway, spend cap, and backpressure semaphore that wraps Claude calls.
- AGNT's tool executor normalizes all tool-calling schemas to a common shape. OpenAI function-calling definitions translate 1:1, so tools authored for GPT work in AGNT without rewriting.
Integration recipes
Prompts & playbooks
Links
Share as social post
AGNT + OpenAI API: GPT models + function calling — AGNT's llm_gateway supports OpenAI as a provider. https://agntdot.com/stack/openai-api
138 / 280 chars