Skip to content

feat: add Avian model provider#2246

Open
avianion wants to merge 2 commits intolanggenius:mainfrom
avianion:feat/add-avian-provider
Open

feat: add Avian model provider#2246
avianion wants to merge 2 commits intolanggenius:mainfrom
avianion:feat/add-avian-provider

Conversation

@avianion
Copy link
Copy Markdown

@avianion avianion commented Apr 4, 2026

Overview

Add Avian as a third-party model provider plugin for Dify.

Avian is an OpenAI-compatible inference API at https://api.avian.io/v1 offering access to leading open-source models at competitive prices.

Models Included

Model Context Window Input Price Output Price
DeepSeek V3.2 164K $0.14/M tokens $0.28/M tokens
Kimi K2.5 128K $0.14/M tokens $0.28/M tokens
GLM-5 128K $0.25/M tokens $0.50/M tokens
MiniMax M2.5 1M $0.15/M tokens $0.30/M tokens

Plugin Details

  • Type: Model provider (LLM)
  • Configuration: Predefined models + customizable model support
  • Auth: API key
  • Privacy policy: Included as PRIVACY.md inside the plugin package

Checklist

  • Packaged as .difypkg
  • PRIVACY.md included in package
  • manifest.yaml with privacy: PRIVACY.md field
  • README.md with contact info and repository URL
  • Follows existing plugin structure conventions
  • Placed in avian/avian/ directory per repo conventions

Contact

Add Avian as a third-party model provider plugin. Avian is an
OpenAI-compatible inference API (https://api.avian.io/v1) offering:

- DeepSeek V3.2 (164K context) - $0.14/$0.28 per M tokens
- Kimi K2.5 (128K context) - $0.14/$0.28 per M tokens
- GLM-5 (128K context) - $0.25/$0.50 per M tokens
- MiniMax M2.5 (1M context) - $0.15/$0.30 per M tokens

Includes PRIVACY.md as required for third-party plugins.
Gmasterzhangxinyang pushed a commit to Gmasterzhangxinyang/dify-plugins that referenced this pull request Apr 6, 2026
Co-authored-by: Yuanbo Li <ybalbert@amazon.com>
@xtaq
Copy link
Copy Markdown

xtaq commented Apr 7, 2026

Nice positioning here — competitive pricing + OpenAI-compatible API + large context windows is a strong combination for developers who want optionality without rewriting their stack.

I’m curious: for a model-provider plugin like Avian, what usually matters more for adoption after listing?

  1. being easy to discover when users compare providers
  2. communicating price/performance tradeoffs clearly
  3. building trust around stability / compatibility / support

Asking because provider plugins seem to win or lose on metadata and presentation almost as much as on the model itself.

@avianion
Copy link
Copy Markdown
Author

avianion commented Apr 7, 2026

Addressed feedback: added credential schemas, sorted models by release date, added tool-call features

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants