itsajchan/openai-agents-sdk-with-fireworks — explained in plain English
Analysis updated 2026-05-18
Learn how to point OpenAI's Agents SDK at a non-OpenAI model provider like Fireworks.
See a minimal example of calling Fireworks through an OpenAI-compatible Responses API.
Reuse the workaround for the Agents SDK misreading a model name that contains slashes.
| itsajchan/openai-agents-sdk-with-fireworks | 0c33/agentic-ai | adennng/stock_strategy_lab | |
|---|---|---|---|
| Stars | 14 | 14 | 14 |
| Language | Python | Python | Python |
| Setup difficulty | easy | hard | hard |
| Complexity | 2/5 | 4/5 | 4/5 |
| Audience | developer | developer | researcher |
Figures from each repo's GitHub metadata at analysis time.
Requires a Fireworks AI API key set as an environment variable before running the demos.
This repository is a small demo project showing how to connect OpenAI's Agents SDK, a toolkit for building AI agents that can call functions, to Fireworks AI, a service that hosts large language models and offers them through an interface compatible with OpenAI's own API. It exists mainly as example code, not a full application. Two Python scripts make up the core demonstration. One creates a simple agent called FileLister that has access to a local function letting it list files, with Fireworks providing the actual model reasoning behind that agent. The other script shows a more minimal example, directly calling Fireworks through the Responses API format without wrapping it in an agent at all. A setup script is included to create a Python 3.12 virtual environment and install the required openai-agents package, and a plain HTML tic-tac-toe page is included as an unrelated standalone demo you can open directly in a browser. Before running either Python demo, you need a Fireworks API key, which you either paste directly into the script or, more safely, set as an environment variable so it is never committed to the repository. Both demos are pointed by default at a specific Fireworks-hosted model. The README calls out one specific technical gotcha: the Fireworks model name contains slashes, and if passed to the Agents SDK the normal way, the SDK misreads part of it as a provider prefix and throws an error. The demo works around this by building an explicit client object pointed at the Fireworks endpoint instead of passing the raw model string. The author reminds users not to commit real API keys, and to rotate any key immediately if one is ever accidentally saved into the project folder. No license is stated in the README.
A small demo showing how to use OpenAI's Agents SDK and Responses API with Fireworks AI as the model provider instead of OpenAI itself.
Mainly Python. The stack also includes Python, OpenAI Agents SDK, Fireworks AI.
No license information is provided in the README.
Setup difficulty is rated easy, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.