Parallel AI Node for n8n

This package provides a custom n8n node to interact with the
Parallel AI platform. The entire node surface —
resources, operations and parameter fields — is auto-generated from the
official Parallel AI OpenAPI spec, so every public endpoint of the API is
available directly in the editor.

Node

The package registers a single Parallel AI node (parallelAi). It follows
the standard n8n Resource / Operation pattern: pick a Resource, then an
Operation, then fill in the parameters for that operation.

The Resource dropdown exposes every category of the Parallel AI API
(Agents, Audio, Brand Kit, Browser, Canvas, Chat, Companies, Company Users,
Content, Content Research, Departments, Documents, Employees, Images, Inbox,
Integrations, Keys, Landing Pages, Leads, Lists, Marketing Campaigns, Mcp,
Memories, Models, Monitors, Organizations, People, Projects, Scheduled Tasks,
SEO, Sequences, Skills, Support, Tags, Usage, Video, Voices, Web, Websites …).
The operations and parameter fields for each resource are generated from the
OpenAPI spec: path/query/body parameters become typed input fields (string,
number, boolean, dropdown or JSON), and required parameters are marked.

The node is marked usable as a tool, so it can be attached directly to an
n8n AI Agent.

v0.9.0 breaking changes

The hand-crafted resources of earlier versions (Browser Task, Document,
Employee, Folder, Image, Knowledge Base, List, Sequence, System, Video) and
the generic API (Any Operation) resource were replaced by the
spec-generated resources above. Workflows built on v0.8 need rebuilding:

• Resource and operation names changed (e.g. Employee → Employees, Chat →
  Chat Employee), and parameter fields are now plain node fields instead of a
  resource-mapper section.
• Convenience behaviors were removed: the browser-task polling loop, the
  knowledge-base result shaping for AI retrievers, and the employee-chat
  output formatting. The node now returns the raw API response of each
  endpoint; to run long-running tasks, call the create/status endpoints in
  sequence (e.g. with a Wait node between them).

Credentials

The node authenticates with a Parallel AI API credential (parallelAiApi).

1. In Parallel AI, open your account settings and copy your API key.
2. In n8n, go to Credentials > New, search for Parallel AI API.
3. Paste the API key. Leave the base URL at its default unless you use a
   self-hosted instance.

Installation

In n8n (community node)

In your n8n instance, go to Settings > Community Nodes, select Install,
and enter the package name:

@parallel-ai/n8n-nodes-parallel-ai

Usage examples

Example 1: Chat with an AI employee

Ask one of your Parallel AI employees a question and use the answer downstream.

1. Add a Manual Trigger node.
2. Add the Parallel AI node and connect it to the trigger.
3. Select your Parallel AI API credential.
4. Set Resource to Employees and Operation to Chat Employee.
5. Fill in the required fields (for example Employee ID and Message,
   or use an expression such as {{ $json.question }} to pass in data from a
   previous node).
6. Execute the node. The output is the raw JSON chat response.

Example 2: Run a browser task

Have a Parallel AI browser agent complete a task described in plain English.

1. Add the Parallel AI node.
2. Set Resource to Browser and choose the task-submission operation
   (e.g. Create Browser Task).
3. In Task, describe what to do, e.g.
   Go to https://news.ycombinator.com and return the titles of the top 5 stories.
4. Execute the node. The response contains a taskId you can pass to the
   task-status operation (after a Wait node) to fetch the result.

Example 3: List resources with pagination

Most list endpoints accept optional pagination fields.

1. Add the Parallel AI node.
2. Set Resource to Companies and Operation to Get Companies.
3. Optionally set Limit and Page.
4. Execute the node. Each returned object becomes one n8n output item.

Local development

npm install
npm run build      # clean, compile TypeScript, copy icons
npm run dev        # tsc --watch
npm run lint       # eslint (includes eslint-plugin-n8n-nodes-base)
npm run format     # prettier
npm run generate   # regenerate nodes/ParallelAi/generated/ from the live OpenAPI spec
node scripts/check-generated.js   # spot-check the built output

Regenerating the node surface

The node's resources, operations and parameter fields are compiled from the
Parallel AI OpenAPI spec by scripts/generate_node.py into
nodes/ParallelAi/generated/ (committed to the repo, so consumers never need
Python). To pick up new or changed endpoints:

npm run generate   # or: python3 scripts/generate_node.py path/to/openapi.json
npm run build && npm run lint && node scripts/check-generated.js

Review the diff, bump the version and publish a new release.

Then link the module into your local n8n installation:

cd /path/to/n8n
npm link /path/to/n8n-nodes-parallel-ai

License

MIT