Home · Agents · Hermes

Use Hermes with Manifest

— / 30 days
Personal AI Agent

Hermes is a lightweight agent for delivering AI capabilities to your stack. Point it at Manifest's endpoint and every request routes through the cheapest capable model. Same Hermes config, smaller bill.

— models routed through Manifest over the last 30 days.

Why use Hermes with Manifest

Hermes exposes a clean YAML-driven config for plugging in any OpenAI-compatible endpoint. Manifest is exactly that: a single endpoint that fans your traffic out to many providers, picks the right model per request, and surfaces cost in one dashboard.

  • One config.yaml change, zero workflow disruption
  • Save up to 70% on a typical Hermes workload
  • Switch between Claude, GPT, Gemini, and open models without rewriting prompts
  • Per-agent spend limits and rate caps
  • Real-time cost dashboards with token, model, and latency breakdowns
  • Open source, self-hostable, OpenAI-compatible

Setup in 60 seconds

Add the Manifest endpoint to your Hermes config.yaml or run hermes model and pick "Custom endpoint".

Point Hermes at the Manifest endpoint to route requests across multiple models.

Open the Hermes configuration file:

hermes config edit

Add the following model: section to your config.yaml: 

model:
  provider: custom
  base_url: https://api.manifest.build/v1
  api_key: mnfst_YOUR_KEY
  default: auto
provider custom
base_url https://api.manifest.build/v1
api_key mnfst_YOUR_KEY
default auto

Hermes tokens processed on Manifest

Models in use

# Model
 Auth type Tokens / last 30 days