hermes-runcrate

A Hermes Agent model-provider plugin for Runcrate.

Hermes runs its own agent loop against Runcrate's OpenAI-compatible API at https://api.runcrate.ai/v1. Open models are free with your GitHub account: DeepSeek V4.1 Flash, GLM-5.3, GLM-5.3 Flash, gpt-oss 120B and Qwen3.8 Flash. Every other model in the catalog is billed per token from a prepaid balance.

Setup

hermes plugins install runcrate-provider && export RUNCRATE_API_KEY=rc_live_... && hermes chat --provider runcrate -m deepseek-ai/DeepSeek-V4.1-Flash

Get the key at https://www.runcrate.ai/dashboard/api-keys (sign in with GitHub). Put it in ~/.hermes/.env to keep it; the installer also asks for it.

Before the catalog entry is merged, install from GitHub instead:

hermes plugins install runcrateai/hermes-runcrate/model-providers/runcrate

Then hermes model and pick Runcrate.

What it does

Provider id	runcrate (aliases runcrate-ai, runcrate.ai)
Base URL	https://api.runcrate.ai/v1 (override with RUNCRATE_BASE_URL)
Auth	RUNCRATE_API_KEY, sent as Authorization: Bearer
Model list	live from GET https://api.runcrate.ai/v1/models; if that fails, the six open models below
Default model	deepseek-ai/DeepSeek-V4.1-Flash (also the auxiliary model)
Reasoning	Hermes' reasoning setting is sent as reasoning_effort; turning reasoning off sends none

Fallback models: deepseek-ai/DeepSeek-V4.1-Flash, deepseek-ai/DeepSeek-V4.1-Flash-1017, zai-org/GLM-5.3, zai-org/GLM-5.3-Flash, Qwen/Qwen3.8-Flash, openai/gpt-oss-120b.

The plugin registers one provider profile and nothing else: no tools, hooks, middleware or extra dependencies. Prompts, tool schemas and the key go to api.runcrate.ai. Free requests can be served by third-party model providers, who handle prompts and outputs under their own terms; see the privacy policy: https://www.runcrate.ai/privacy

Check it

hermes plugins validate --install-deps model-providers/runcrate

License

MIT, Aeonmind LLC.