openhandle

n8n node for Openhandle: read public Instagram, TikTok, X, and Reddit data

Package Information

Released: this month
Downloads: 0 weekly / 0 monthly
Latest Version: 0.1.0
Author: Openhandle

Available Nodes

Documentation generating…

Documentation

@openhandle/n8n-nodes-openhandle

This is an n8n community node for Openhandle. It reads public Instagram, TikTok, X, and Reddit data inside your n8n workflows.

Look up profiles or pull an account's latest posts. You can also search hashtags. Responses use the same top-level JSON fields across all four platforms. The node also works as a tool for the n8n AI Agent.

Install

In n8n, open Settings > Community Nodes, select Install, and enter @openhandle/n8n-nodes-openhandle.

The n8n docs explain how community nodes work.

Credentials

  1. Create a free account at app.openhandle.dev/signup. You get 100 free requests, and no card is needed.
  2. Open API Keys in the dashboard and create a key.
  3. In n8n, add an Openhandle API credential and paste the key.

Keys that start with oh_live_ read real public data. Keys that start with oh_test_ return free synthetic data. A test key is a good way to build a workflow before you spend anything. Try the Instagram Get Profile operation with @northstar_forge_test.

Resources

Resource Operations What you can read
Instagram 43 Profiles, posts, reels, stories, highlights, comments, hashtags, locations, music, search
TikTok 28 Profiles, posts, comments, playlists, hashtags, effects, music, trending lists, search
X 15 Profiles, posts, replies, reposters, lists, search
Reddit 20 Profiles, posts, comments, subreddits, wiki pages, trending posts, search
URL 1 Paste a supported post or profile link and get the same data as the matching operation

Profiles take a username with an @ prefix, such as @openai, or the platform ID. Posts take the platform post ID or shortcode. If you only have a link, use the URL resource.

Output

List and search operations return one n8n item for each result. Get operations return one item with the full response: platform, resource, capturedAt, source, data, and meta.

List operations read the first page by default. Each page is a separate billed request. Keep Max Pages low when testing with a live key. Turn on Return All to read more pages, up to Max Pages (5 by default).

Freshness and pricing

Every operation has a Freshness field. It sets how old the data may be. Older cached answers cost less.

Freshness Price per answered request
Live $0.0025
24 Hours (default) $0.0005
7 Days $0.0001
30 Days $0

The source field in each response tells you if the answer came from the platform just now (live) or from the cache (cache).

Links

License

MIT

Discussion