providers
Together
A broad, reliable open-model menu (GLM, DeepSeek, Llama, gpt-oss) with dedicated-capacity options when you outgrow shared.
| endpoint | https://api.together.xyz/v1 |
| env var | TOGETHER_API_KEY |
| get a key | api.together.xyz |
Get an API key
- Create an account at api.together.xyz
- Open Settings → API Keys
- Click 'Create key' and name it
- Copy the key value (shown only once)
- Set TOGETHER_API_KEY
Small starter credit for new accounts; ~$5 top-up for continued use.
Enable
export TOGETHER_API_KEY=... # shell profile — survives foreveror in the terminal — validates the key live before saving:
/enable togetheror pick any locked 🔒 Together model in /models and paste the key inline.
Models
Starter menu (once enabled, /models fetches the provider's full live list):
openai/gpt-oss-120bzai-org/GLM-4.7deepseek-ai/DeepSeek-V3meta-llama/Llama-3.3-70B-Instruct-Turbo
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://api.together.xyz/v1", # key read from $TOGETHER_API_KEY
)Notes
- Ids are HF-style (
zai-org/GLM-4.7,deepseek-ai/DeepSeek-V3).