Skip to main content

Configure Hermes Agent with KiosAPI

Hermes Agent can use KiosAPI through an OpenAI-compatible custom provider. The exact configuration fields may change between Hermes releases, so use the current Hermes setup or provider configuration interface and apply the values below.

Required values

Create the token in the group matching your workflow — Codex-Pro/Codex-Plus for gpt-5.x/gpt-6 models, Claude-Kiro for Claude models, Gemini for Gemini, or Free for the free tier (after Telegram verification). See Choosing a Token Group. A typical Hermes custom provider block in config.yaml:

Verify the model list

Choose a model intended for tool use and coding workflows. Availability and pricing are shown on the live Models & Pricing page.

Configure the provider

  1. Open Hermes provider/model configuration.
  2. Add a custom OpenAI-compatible provider.
  3. Enter https://kiosapi.com/v1 as the base URL.
  4. Store the KiosAPI key in the provider credential field or configured environment variable.
  5. Add the selected model ID and make it active for the target Hermes profile.
  6. Restart the affected Hermes gateway or session if the configuration interface requires it.

Test the connection

Start a new Hermes session and send a short prompt that does not require tools. If it fails, compare the configured model ID with GET /v1/models, then run the cURL test.
Do not place an API key in a committed config file or share a full Hermes diagnostic log without sanitizing secrets.

Common problems

  • 401 Unauthorized: key missing, revoked, or stored under the wrong provider.
  • 404 Not Found: invalid model ID or incorrect base URL.
  • /v1/v1/...: Hermes/provider adapter appended /v1; set the configured host to https://kiosapi.com.
  • Model responds but tools fail: select a model that supports tool calling and verify the Hermes tool configuration separately.