Skip to content

feat(components): add GreenPT provider - #6674

Open
robertkeus wants to merge 1 commit into
FlowiseAI:mainfrom
robertkeus:codex/add-greenpt-provider
Open

feat(components): add GreenPT provider#6674
robertkeus wants to merge 1 commit into
FlowiseAI:mainfrom
robertkeus:codex/add-greenpt-provider

Conversation

@robertkeus

Copy link
Copy Markdown

What changed

This PR adds first-class GreenPT support across the Flowise capabilities used in production RAG and agent workflows.

GreenPT is a European AI provider with an OpenAI-compatible API, optimized infrastructure, and data centers powered by 100% renewable energy. The native integration gives Flowise users a shared GreenPT credential and provider-specific behavior that the generic OpenAI-compatible nodes do not cover.

Capabilities

  • Chat model node with authenticated live model discovery from GET /v1/models
    • defaults to the flagship glm-5.2 model
    • highlights kimi-k2.7-code for coding workflows
  • Embedding node with live model discovery and green-embedding as the default
  • Rerank retriever using green-rerank and the native POST /v1/rerank endpoint
  • Speech-to-text settings for GreenS Pro and GreenS using the native Deepgram-compatible POST /v1/listen protocol
  • One password-type GreenPT API credential shared by all four capabilities
  • Official GreenPT mark for component and speech settings UI

Why a native integration

Flowise can already call OpenAI-compatible chat and embedding endpoints through custom nodes, but that still requires users to enter model IDs and endpoint details manually. It also does not provide GreenPT reranking, the GreenPT speech protocol, or a single provider credential across those workflows.

This integration keeps the standard LangChain interfaces for chat and embeddings while adding live discovery and the native GreenPT endpoints where compatibility differs.

Reliability and security

  • Model discovery uses the fixed GreenPT production endpoint and a bounded timeout.
  • The API key is stored as a password credential.
  • Speech requests validate the content-type header and never expose the key in errors.
  • Speech and rerank responses are shape-validated before use.
  • Reranking returns new Document objects instead of mutating retrieved inputs.

Validation

  • Full components Jest suite: 24 suites, 922 tests passed
  • Focused GreenPT suite: 12 tests passed
  • Flowise components TypeScript and Gulp build passed
  • Flowise UI production Vite build passed
  • ESLint passed for every changed file
  • git diff --check passed

The repository-wide components lint command currently reports nine pre-existing unused eslint-disable directives in untouched files. No new dependencies are introduced.

@gemini-code-assist

Copy link
Copy Markdown
Contributor

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

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.

1 participant