n8n-nodes-canny

This is an n8n community node. It lets you use Canny in your n8n workflows.

Canny is a product feedback and feature-request management tool. It lets teams collect feedback from users, organize it into boards and categories, track votes and comments, and publish public roadmaps.

n8n is a fair-code licensed workflow automation platform.

Installation
Operations
Credentials
Compatibility
Usage
Resources
Version history

Installation

Follow the installation guide in the n8n community nodes documentation.

Operations

Post

• Create
• Get
• Get Many
• Update
• Delete
• Change Status

Board

• Get
• Get Many

Category

• Get
• Get Many
• Create
• Delete

Idea

• Get
• Get Many
• Delete

Portal Comment

• Get
• Get Many
• Create
• Delete

User

• Get Many

Canny Trigger

Starts a workflow when a Canny event occurs (post created, status changed, comment created, vote cast, and more). Requires manual webhook setup — see Usage below.

Not yet supported: Idea filtering/sort/Merge, Post's customFields/imageURLs on Create and Update. These may be added in a future version.

Credentials

You'll need a Canny account with API access:

1. Sign up for a free Canny account at canny.io if you don't already have one.
2. In your Canny dashboard, go to Settings > API.
3. Copy your Secret API Key.
4. In n8n, create a new Canny API credential and paste the key into the API Key field.

Canny's API expects the key as a field in the request body rather than a header — this credential handles that automatically.

Compatibility

Tested against n8n using n8n-workflow's programmatic node style (@n8n/node-cli). No known version incompatibilities.

Usage

Board, User, and Post fields

Board, User (Author/Changer), and Post ID fields use n8n's resource locator: pick from a searchable list, or switch to "By ID" and paste a raw Canny ID directly.

Pagination

Canny's API has two different pagination styles depending on the resource:

• Post, Category use skip-based pagination (skip/limit/hasMore).
• Portal Comment, Idea, User use cursor-based pagination (cursor/hasNextPage).

This node handles both automatically — just use the Return All and Limit fields on any "Get Many" operation.

Board has no server-side pagination at all; "Get Many" always returns your full list of boards in a single request.

Setting up the Canny Trigger

Canny has no API for registering webhooks, so setup is manual:

1. Add the Canny Trigger node to your workflow, choose your Canny credential, and select the Event you want to trigger on.
2. Copy the node's Webhook URL (shown in the node once activated).
3. In your Canny dashboard, go to Settings > API > Webhooks, and paste the URL there.
4. Canny will send every subscribed event to that same URL — this node automatically verifies each payload's signature and filters out any event that doesn't match your selected Event, so only the one you configured actually starts the workflow.

Using this node with AI Agents

This node has usableAsTool: true and works as a tool for n8n's AI Agent node. A few things to know:

• Each operation needs its own dedicated tool node. Don't wire a single generic Canny tool node into the Agent and expect the model to pick Resource/Operation freely — set the Resource and Operation explicitly on each tool node (e.g. one node configured as "Post → Change Status", a separate node as "User → Get Many"), then connect each to the Agent's Tool input separately.
• Leave ID fields set to "Defined automatically by the model." For any field that's a resource locator (Board, User, Post), the AI tool interface only supports "By ID" mode — the interactive "From list" picker isn't available to a model, since browsing a dropdown requires human interaction. The agent must supply a real Canny ID as a plain string.
• Give the agent a way to discover real IDs first. Since the model can't browse pickers, connect a "Get Many" tool for whichever resource it needs an ID for (e.g. a "User → Get Many" tool so it can look up a user's ID before calling an operation that requires one). Without that, the agent has no way to know a valid ID and may either ask you for it or hallucinate a plausible-looking one.

For example, to let an agent change a post's status, you'd typically connect three separate tool nodes: one for Post → Get Many (to find the post), one for User → Get Many (to find the user), and one for Post → Change Status (with postID, changerID, status, and commentValue all left as model-defined parameters).

Whether the model actually chains multiple tool calls together on its own (e.g. looking up a user ID before using it) depends on the underlying LLM's own reasoning — this node provides everything needed for that chain to work, but can't force a model to take the extra discovery step. If an agent asks you for an ID instead of looking it up itself, try being more explicit in your prompt about the order of steps.

For general help getting started with n8n, see the Try it out documentation.

Resources

• n8n community nodes documentation
• Canny API reference

Disclaimer

This is an unofficial, community-built integration. It is not affiliated with, endorsed by, or supported by Canny Inc. "Canny" is a trademark of Canny Inc., used here under nominative fair use for descriptive purposes only.

Version history

0.2.13

Renamed the package from the unscoped n8n-nodes-canny to the scoped @automiq/n8n-nodes-canny, to reduce trademark-dispute risk on the unscoped npm namespace. The old package is now deprecated.

0.2.10 – 0.2.12

Fixed AI agent tool-orchestration gaps by making the Resource field's description explicitly enumerate valid values. Moved AI-agent "By ID" hints into the shared resourceLocatorField helper to reduce duplication. Excluded description files from strict duplication analysis (structural repetition inherent to n8n's declarative field style). Documented a known limitation around multi-step AI agent tool chaining.

0.2.9

Added a User resource (Get Many) so AI agents can discover real Canny user IDs. Added resource-locator pickers (searchable dropdowns) for Board, User, and Post ID fields across all resources, replacing raw text ID entry. Improved API error messages to include the failing endpoint and HTTP status code. Added missing field placeholders throughout. Clarified resource-locator field descriptions for AI agent usage. Documented AI Agent tool-wiring patterns.

0.2.6 – 0.2.7

Reduced the execute() method's cognitive complexity by extracting pagination-dispatch and error-classification logic into dedicated helper functions. Consolidated repeated Post ID field definitions into the shared idField helper, further reducing code duplication.

0.2.5

Removed Board's non-functional Return All/Limit fields (Canny's boards/list endpoint doesn't support pagination). Normalized Return All defaults and Limit bounds across Post, Category, Idea, and Portal Comment. Migrated Board and Category onto the shared field-definition helpers for consistency with the rest of the codebase.

0.2.4

Extracted shared icon, credential, and dropdown-field helpers to reduce code duplication across resource description files. Added Jest unit test coverage for all request builders (Post, Board, Category, Idea, Portal Comment) and for the Canny Trigger's webhook signature verification.

0.2.3

Added Post's Update, Delete, and Change Status operations. Implemented real webhook signature verification (HMAC-SHA256) and event filtering in Canny Trigger. Fixed a bug where Board's Get Many sent undocumented pagination parameters that silently broke the response. Fixed Portal Comment's Get Many using the wrong API version (v1 instead of v2). Centralized icon files into a shared folder.

0.2.0 – 0.2.2

Internal refactoring: extracted shared field-definition helpers to reduce duplication across resource description files, simplified build tooling to use the official @n8n/node-cli scaffold consistently.

0.1.0

Initial release. Supports Post (Create/Get/Get Many), Board (Get/Get Many), Category (Get/Get Many/Create/Delete), Idea (Get/Get Many/Delete), and Portal Comment (Get/Get Many/Create/Delete). Canny Trigger node included as a manual-setup webhook receiver — payload signature verification not yet implemented.