providers
OpenRouter
One key, essentially every model on the market — plus :free variants of many open models. The best single-key starting point.
| endpoint | https://openrouter.ai/api/v1 |
| env var | OPENROUTER_API_KEY |
| get a key | openrouter.ai |
Get an API key
- Sign up at openrouter.ai (Google or email)
- Open the Keys page (openrouter.ai/settings/keys)
- Click 'Create Key' and name it
- Copy the key now — it's shown only once
- Set OPENROUTER_API_KEY
Free (:free) models work at $0 balance; add credits for paid models/limits.
Enable
export OPENROUTER_API_KEY=... # shell profile — survives foreveror in the terminal — validates the key live before saving:
/enable openrouteror pick any locked 🔒 OpenRouter model in /models and paste the key inline.
Models
Starter menu (once enabled, /models fetches the provider's full live list):
openai/gpt-oss-120bz-ai/glm-4.7moonshotai/kimi-k2deepseek/deepseek-chatqwen/qwen3-235b-a22b
Switch anytime — /model gpt-oss-120b fuzzy-matches; context carries over.
SDK
from mantis_agent import MantisAgentOptions
options = MantisAgentOptions(
model="openai/gpt-oss-120b",
backend="https://openrouter.ai/api/v1", # key read from $OPENROUTER_API_KEY
)Notes
- Try
:freemodel variants (e.g.meta-llama/llama-3.3-70b-instruct:free) before spending anything. - Model ids are namespaced (
z-ai/glm-4.7,moonshotai/kimi-k2).