Lemmy Instance Extractor
Pricing
from $0.35 / 1,000 items
Lemmy Instance Extractor
Point at ANY Lemmy instance (lemmy.world, sh.itjust.works, lemm.ee…) and pull instance stats, posts, comments or communities into clean flat rows. One actor reads every federated /api/v3 server. Sort + community filters, auto-paged, lossless _raw. Pay per item.
Pricing
from $0.35 / 1,000 items
Rating
0.0
(0)
Developer
Datamule
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
23 days ago
Last modified
Categories
Share
Point at ANY Lemmy instance and pull its stats, posts, comments, or communities into clean, flat, tabular rows — one actor for the entire federated network.
Lemmy is the open-source, federated link-aggregator (a self-hosted Reddit alternative). Every Lemmy server — lemmy.world, sh.itjust.works, lemm.ee, programming.dev, and thousands of others — exposes the identical, unauthenticated /api/v3 REST API. This actor speaks that one grammar, so a single run can read across as many instances as you give it and return comparable rows for every one.
No login, no API key, no browser automation — just clean JSON.
What it does
Give it a list of instances and a mode, and it returns one flat row per item:
| Mode | Endpoint | One row per… | Use it for |
|---|---|---|---|
site | /api/v3/site | instance | A federated-instance census: name, version, registration policy, total users/posts/comments/communities, and daily/weekly/monthly/half-year active users. |
posts | /api/v3/post/list | post | Front-page or per-community post feeds — title, URL, body, author, community, score, up/downvotes, comment count. |
comments | /api/v3/comment/list | comment | Comment feeds — content, author, post, community, score, reply count. |
communities | /api/v3/community/list | community | A community directory — name, title, description, subscribers, post/comment counts, active users. |
Every row also carries the lossless original object in _raw, plus _instance / _mode / _sort provenance so multi-instance, multi-mode datasets stay unambiguous.
Why this instead of a single-instance scraper
Most Lemmy scrapers target one instance or one community. This one is generic: the same run censuses lemmy.world and sh.itjust.works and lemm.ee side by side, because they all speak the same API. Point it at a fresh instance you've never seen and it just works — the federated-network moat a single-target scraper can't match.
Input
| Field | Type | Description |
|---|---|---|
instances | array (required) | Instances as bare hostnames or base URLs, e.g. ["lemmy.world", "https://sh.itjust.works"]. |
mode | string | site (default), posts, comments, or communities. |
sort | string | Feed sort order (ignored for site). Active, Hot, New, Old, MostComments, TopDay…TopAll, Controversial, Scaled. |
limit | integer | Max items per instance in feed modes. Lemmy caps a request at 50, so larger values are auto-paged. Default 50. |
communityName | string | Optional. Restrict posts/comments to a community — technology or technology@lemmy.world. |
communityId | integer | Optional. Restrict posts/comments to a community by numeric ID (takes precedence over name). |
postId | integer | Optional. In comments mode, restrict to one post's thread. |
listingType | string | All (federated, default) or Local (this instance only) for posts/communities. |
userAgent / extraHeaders | string / object | Optional. Not required for public instances. |
Example — a two-instance census
{"instances": ["lemmy.world", "sh.itjust.works"],"mode": "site"}
Example — top posts of the week from one community
{"instances": ["lemmy.world"],"mode": "posts","sort": "TopWeek","communityName": "technology","limit": 100}
Notes on Lemmy's API (handled for you)
limitmaxes out at 50 per request — this actor pages the 1-indexedpageparam automatically to reach your totallimit.- The comment feed supports fewer sorts than posts (
Hot/Top/New/Old/Controversial). If you pick a sort a mode doesn't support, it's mapped to the closest valid one — your run never fails on an unsupported sort. - Resilient by instance: an instance that's down, rate-limited, or isn't a Lemmy server is skipped with a warning and the run continues. If every instance fails, the run errors out rather than returning a misleading empty result.
Pricing
Pay-per-result: you're charged one item event per row emitted (one per instance in site mode; one per post/comment/community in the feed modes). No monthly rental.
If you set a maximum cost per run, the Actor delivers only the rows that budget covers: every row in the dataset is billed, nothing unpaid is written, and no further page or instance is requested once the budget cannot pay for another row. If a later page fails after some rows were delivered, those rows stay (and are billed) and the run ends as failed so you can see it was incomplete.
Output row (fields)
Provenance _source / _instance / _mode / _sort; identity id / name / title / body / url / ap_id / published / updated; flags local / nsfw / removed / deleted / locked; actors creator_id / creator_name / creator_actor_id; relations community_* / post_id / post_name; item counts score / upvotes / downvotes / num_comments / child_count; container counts subscribers / subscribers_local / num_posts / num_communities / users / active_day|week|month|half_year; instance instance_version / registration_mode / admins_count; and the lossless _raw. Every field is nullable — a mode fills its subset and leaves the rest empty.