hermes - 💡(How to fix) Fix Feature Request: Add official ByteDance / BytePlus ModelArk provider

Feature Request: Add official ByteDance / BytePlus ModelArk provider

Provider name: ByteDance / BytePlus ModelArk Official website: https://www.byteplus.com/en/product/modelark API compatibility: Full OpenAI compatible, full Anthropic compatible endpoints available.

Description

Currently ByteDance ModelArk provider is missing from the official hermes model selector. This is now the single largest LLM provider that is not natively supported in Hermes.

As of today ModelArk exposes 48 official production models including:

• seed-2-0-pro-260328
• deepseek-r1 / deepseek-v4
• kimi-k2
• glm-4-7
• All Skylark / Seedream / Seedance models
• All image / video / 3D generation models

Right now users can configure it manually as a custom provider, but having it officially supported will remove a lot of friction for new users.

Technical details

✅ Use these official global endpoints, do NOT use the China internal volces endpoint, it will incur extra cost and require China billing:

• Authentication: Standard Bearer API key
• Full native support for: streaming, tool calling, vision, structured outputs, reasoning traces
• Official documentation: https://docs.byteplus.com/en/docs/ModelArk/1928261

Required changes

This provider is 100% API compatible. No implementation changes are needed. Only need to:

1. Add entry to providers list in hermes_cli/config.py DEFAULT_CONFIG
2. Add to the interactive hermes model selector menu
3. Add to public providers documentation page

This is currently the most requested missing provider. Hundreds of users are already configuring this manually today via custom endpoints, adding it officially would be a very high impact change with zero maintenance overhead.

This single provider addition will give Hermes users native access to every major model currently available on the market, in one click.