Fireworks AI

Run Browser Use agents on Fireworks-served Kimi K3 and GLM 5.3 Flash, in Browser Use Cloud or with your own Fireworks key in the open-source library.

Fireworks AI serves open models with fast inference. You can use it with Browser Use in two ways:

• Browser Use Cloud runs kimi-k3 and glm-5.3-flash on Fireworks. Pass the model ID; you need no Fireworks account.
• The open-source library calls any Fireworks model with your own FIREWORKS_API_KEY through its OpenAI-compatible endpoint.

Browser Use Cloud

Model	API string	Notes
Kimi K3	kimi-k3	Moonshot’s K3 on Fireworks’ fast serving path. Vision and long context.
GLM 5.3 Flash	glm-5.3-flash	Z.ai’s first multimodal GLM 5 model. Low cost per step.

Get a key from Browser Use Cloud and pass the model when you create a run:

Python

from browser_use_sdk.v4 import BrowserUse

client = BrowserUse()
run = client.runs.create(
    "Open https://news.ycombinator.com and return the title of the top story.",
    model="kimi-k3",
)
result = client.runs.wait_for_completion(run.id)
print(result.result)

Usage is billed to your Browser Use credits. These models do not take a bring-your-own Fireworks key in Cloud. Free-plan projects need to buy credits before using them. See Models for every V4 model ID.

Open-source library

Install Browser Use, get a key from the Fireworks dashboard, and point ChatOpenAI at the Fireworks endpoint:

import asyncio
import os

from browser_use import Agent, ChatOpenAI
from dotenv import load_dotenv

load_dotenv()

llm = ChatOpenAI(
    model="accounts/fireworks/models/kimi-k3",
    base_url="https://api.fireworks.ai/inference/v1",
    api_key=os.getenv("FIREWORKS_API_KEY"),
)


async def main():
    agent = Agent(
        task="Open https://news.ycombinator.com and return the title of the top story.",
        llm=llm,
    )
    await agent.run()


asyncio.run(main())

.env

FIREWORKS_API_KEY=

Any Fireworks model ID works in model, such as accounts/fireworks/models/glm-5p3-flash. Fireworks bills these calls to your Fireworks account.