@aparte/runner-n8n

Put aparté on an n8n chat workflow. The workflow's agent runs in n8n, with its model, its tools and its memory; aparté draws the chat on your page, streamed when the Chat Trigger streams — in place of n8n's own chat widget.

npm install @aparte/runner-n8n @aparte/core

import { AparteClient } from '@aparte/core';
import { createN8nRunner } from '@aparte/runner-n8n';

new AparteClient({
    streamRunner: createN8nRunner({ webhookUrl: 'https://your-n8n.example.com/webhook/<id>/chat' }),
}).start();

• The workflow: a Chat Trigger made publicly available, with your site in its Allowed Origins (CORS). Set its Response Mode to Streaming (and leave streaming on in the AI Agent) to see the answer arrive as it is written, or to When Last Node Finishes for a whole answer.
• Memory is n8n's, per session: each turn sends only the new message. Give sessionId the conversation's id when your app has several, or keeps one across reloads.
• Images the person attaches go to the workflow as files, as the Chat Trigger takes uploads.
• Errors say what to change: an inactive workflow, a trigger that streams while its agent does not, a workflow that answers through response nodes, a proxy that timed out.

No n8n SDK is pulled in: the answer is plain JSON, parsed here, with @aparte/core as the only peer dependency.

Part of the aparté monorepo. ESM-only. See Providers → n8n in the docs.