AICredits logo
Integrations

Flowise

Use AICredits in Flowise chatflows and agentflows through the OpenAI-compatible ChatOpenAI configuration.

Use this page with an AI assistant

Opens a new chat with this docs URL and the correct AICredits base URLs.

Flowise is a visual builder for LLM chatflows, RAG pipelines, and agents. Because AICredits is OpenAI-compatible, Flowise can call AICredits from OpenAI or ChatOpenAI nodes.

ChatOpenAI Node

In a Flowise chatflow or agentflow, add an OpenAI-compatible chat model node and configure it with:

FieldValue
Credential / API keysk-your-aicredits-key
Base path / base URLhttps://api.aicredits.in/v1
Model namePick from Models

Recommended starter model:

openai/gpt-4o-mini

For higher-quality coding, planning, or agent tasks, use:

anthropic/claude-sonnet-4.5
google/gemini-2.5-pro

Flowise field names vary by version. Use the field that maps to OpenAI's base URL, base path, or configuration base path, and set it to https://api.aicredits.in/v1.

Example RAG Flow

Use this basic Flowise structure:

Document Loader → Text Splitter → Vector Store Retriever → Conversational Retrieval QA Chain → ChatOpenAI

Set the ChatOpenAI model to an AICredits model ID. This lets the retrieval and prompt orchestration stay inside Flowise while model access, billing, and usage logs go through AICredits.

Hostinger Deployment Notes

If your Flowise container is internet-facing, enable Flowise authentication and keep it updated. Flowise workflows can call external tools and execute agent steps, so treat a public Flowise instance as a production admin surface.

Use a separate AICredits API key per Flowise environment:

EnvironmentRecommended key
TestingLow monthly cap
Internal teamTeam-specific monthly cap
Production chat appDedicated production key

Verify

Run the chatflow with:

Say "AICredits via Flowise" and nothing else.

Then check Dashboard → Usage for the request.

Troubleshooting

OpenAI key required — Paste the AICredits key in the OpenAI credential field.

Wrong endpoint — The base URL must include /v1.

Model list is empty — Type the model ID manually from Models.

On this page