Swap to a cloud provider

The same Calculator agent from Track 1 runs on a hosted provider. Provider, model, and API key all live in .env so switching providers is a config change, not a code change.

1. Configure#

Copy the template and fill in your key:

bash

cp .env.example .env

.env sets which provider and model to use, plus the matching API key:

bash

LLM_PROVIDER=fireworks
LLM_MODEL=accounts/fireworks/models/gpt-oss-120b
FIREWORKS_API_KEY=fw_...

main.py reads all three from the environment. The calculator.py agent and its primitives are unchanged from Track 1.

2. Run it#

bash

uv run main.py

Output:

text

3. Switch providers#

Edit .env. To use OpenAI instead of Fireworks:

diff

- LLM_PROVIDER=fireworks
- LLM_MODEL=accounts/fireworks/models/gpt-oss-120b
+ LLM_PROVIDER=openai
+ LLM_MODEL=gpt-4.1-nano

Provider	`LLM_PROVIDER`	Example `LLM_MODEL`	Key env var
OpenAI	`openai`	`gpt-4.1-nano`	`OPENAI_API_KEY`
Anthropic	`anthropic`	`claude-haiku-4-5`	`ANTHROPIC_API_KEY`
Groq	`groq`	`llama-3.3-70b-versatile`	`GROQ_API_KEY`
Fireworks	`fireworks`	`accounts/fireworks/models/gpt-oss-120b`	`FIREWORKS_API_KEY`

Model names change over time, so check the provider's own docs for current options.

Takeaway#

Local or cloud, Fireworks or OpenAI, the choice is configuration. Your primitives, plan logic, and task stay put.