All integration research

Research archive · May 2026. Provider access, prices, and terms may have changed. These reports explore potential integrations; they do not establish product availability.

Social Platforms, Community Sites, Forums, Q&A, Messaging, Decentralized Social

38 sources. Scores: Data Richness (DR), Stability Risk (SR — higher = worse), Long-term Viability for Vivly (LTV).

TIER 1: Major Public Social Platforms

1. Reddit (Official Data API)

  • URL: https://www.reddit.com/dev/api / https://developers.reddit.com/docs
  • Exposes: Posts, comments, subreddit metadata, user profiles, voting/score, search, modlogs, listings (hot/new/top/rising). No historical access beyond ~1000 posts per listing.
  • Auth: OAuth 2.0 (script, web, installed app types)
  • Pricing: Free tier (non-commercial, 100 QPM OAuth / 10 QPM unauth). Commercial: ~$12K/mo for 50M calls, ~$0.24 / 1K beyond. Enterprise quotes negotiated.
  • Rate limits: 100 req/min OAuth (commercial-capable apps need a paid tier for production)
  • DR: 5 | SR: 5 (killed third-party apps July 2023, Apollo/RIF dead) | LTV: 3
  • Legal: Strict commercial-use clause; aggressive enforcement (sued Anthropic 2025). No scraping permitted by ToS.
  • Dev friendliness: Good docs, official PRAW (Python), many SDKs
  • Gotchas: No historical depth beyond 1000 items per endpoint. Pushshift was effectively killed for public use in 2023 (now mod-only). Reddit blocks search engines except Google.

2. X / Twitter API v2

  • URL: https://developer.x.com
  • Exposes: Tweets, replies, retweets, quote posts, user metadata, follow graph, lists, bookmarks, search, filtered stream
  • Auth: OAuth 2.0 + Bearer tokens
  • Pricing (Feb 2026 reset): Pay-per-use default — $0.005/read, $0.01/write, $0.20 if post contains URL; cap 2M reads/mo. Legacy Basic $200/mo and Pro $5K/mo (existing subs only). Enterprise $42K-$50K+/mo.
  • Rate limits: Tier-dependent; full-archive search Enterprise only
  • DR: 5 | SR: 5 (most volatile API in industry) | LTV: 2
  • Legal: X v. Bright Data (2024) — court ruled scraping public data without login does NOT violate ToS. X has sued others (Center for Countering Digital Hate, Bright Data lost on appeal partially).
  • Dev friendliness: Docs mediocre; SDKs sparse since v1.1 deprecation
  • Gotchas: Pricing changed 4+ times since 2023. New devs cannot subscribe to Basic/Pro tiers anymore. Free tier write-only (no read).

3. LinkedIn (Official + Unofficial)

  • URL: https://developer.linkedin.com
  • Exposes (official): Member's own profile, posting on behalf of users (Share API), Marketing/Ads, Talent. No third-party profile lookup, no search, no company data lookup.
  • Auth: OAuth 2.0 (Partner Program approval required for most useful scopes)
  • Pricing: Partner-only; most apps rejected
  • Rate limits: Per-product throttles, undisclosed
  • DR: 5 (data) but 1 (what you can access) | SR: 5 | LTV: 1 (official) / 2 (scraping)
  • Legal: hiQ v. LinkedIn settled 2022 — hiQ paid $500K and destroyed data. Proxycurl shut down July 2026 after LinkedIn lawsuit (Jan 2026).
  • Dev friendliness: Docs decent; partner gate is brutal
  • Gotchas: Effectively closed. Scraping carries CFAA + ToS contract risk. Surviving alternatives: Bright Data (legally defended), Apify, LinkdAPI, Apollo.io. Datacenter IPs increasingly blocked.

4. Discord

  • URL: https://discord.com/developers/docs
  • Exposes: Guild messages, members, roles, threads, reactions, voice states, presence — only for guilds your bot is in
  • Auth: Bot tokens, OAuth 2.0 for user data
  • Pricing: Free
  • Rate limits: 50 req/sec global per bot; per-route buckets; 10K invalid req / 10 min IP cap
  • DR: 5 (within accessible guilds) | SR: 2 | LTV: 4
  • Legal: Self-bots/user-tokens against ToS and bannable. Bot must be invited to each guild — no global scraping.
  • Dev friendliness: Excellent — discord.js, discord.py, JDA, detailed docs
  • Gotchas: Message Content Intent is privileged (verified bots must apply, >100 guilds requires Discord approval). No way to read public guild without joining.

5. Slack

  • URL: https://api.slack.com / https://docs.slack.dev
  • Exposes: Channel messages, files, users, conversations, search (paid), workflows
  • Auth: OAuth 2.0, Bot tokens, User tokens
  • Pricing: Free for app creation; Enterprise Grid required for org-wide data
  • Rate limits: Tiered (1, 20, 50, 100 req/min by method). March 2026 change: conversations.history/replies for non-Marketplace apps cut to 1 req/min, 15 objects max.
  • DR: 5 (within authorized workspaces) | SR: 4 (active tightening) | LTV: 3
  • Legal: Workspace owner consent required; SOC2/HIPAA implications
  • Dev friendliness: Excellent docs, Bolt SDK, Block Kit Builder
  • Gotchas: 2026 rate-limit collapse for non-Marketplace apps essentially killed unverified integrations. Must list on Marketplace or get Enterprise Grid contract.

6. Telegram

  • URL: https://core.telegram.org/bots/api (Bot) | https://core.telegram.org/mtproto (MTProto)
  • Exposes (Bot): Messages in chats bot is in, 50MB file cap. (MTProto): Full client capabilities — public channels, search, user accounts.
  • Auth: Bot token (Bot API) / phone+SMS (MTProto)
  • Pricing: Free
  • Rate limits: Bot: 30 msg/sec global, 1 msg/sec per chat, 20 msg/min per group. MTProto: looser but flood-wait enforced.
  • DR: 5 | SR: 2 | LTV: 5
  • Legal: MTProto requires real account (ToS gray area for automation at scale). Public channels via t.me/s/ are SEO-indexable (legitimate scraping).
  • Dev friendliness: Bot API excellent; MTProto via Telethon/Pyrogram/MadelineProto
  • Gotchas: Bot API can't read public channels unless added as admin. MTProto = best path for unified public-channel ingest. Watch for account bans.

7. Mastodon

  • URL: https://docs.joinmastodon.org/api/
  • Exposes: Toots, boosts, favourites, follows, hashtags, lists, public/local/federated timelines, streaming
  • Auth: OAuth 2.0 per-instance
  • Pricing: Free
  • Rate limits: Default 300 req / 5 min per user, 7500 req / 5 min per IP — per instance, configurable
  • DR: 4 | SR: 1 | LTV: 5
  • Legal: AGPL software; per-instance ToS varies. No central entity to sue.
  • Dev friendliness: Mastodon.py, masto.js, well-documented REST
  • Gotchas: Each instance is separate — Vivly needs to discover + auth across hundreds. Streaming API requires read:streaming scope. authorized-fetch mode increasingly common.

8. Bluesky / AT Protocol

  • URL: https://docs.bsky.app / https://atproto.com
  • Exposes: Posts, likes, reposts, follows, lists, profiles. Firehose = full network event stream. Jetstream = JSON-simplified firehose.
  • Auth: App passwords or OAuth (rolling out); firehose is unauthenticated
  • Pricing: Free (public Jetstream instances hosted by Bluesky)
  • Rate limits: XRPC endpoints: ~3000 points/5 min on bsky.social; firehose unlimited
  • DR: 5 | SR: 2 | LTV: 5 (best risk-adjusted bet in social)
  • Legal: Federated, MIT-licensed protocol — minimal vendor lock
  • Dev friendliness: Excellent — @atproto/api, atproto Python, comprehensive docs
  • Gotchas: Run your own Jetstream instance if you need guaranteed availability. CBOR firehose harder than JSON Jetstream.

9. Threads (Meta)

  • URL: https://developers.facebook.com/docs/threads
  • Exposes: Own user's posts, replies, insights, publishing. Limited public discovery.
  • Auth: Meta Graph OAuth + Tech Provider Verification (~1 week review)
  • Pricing: Free
  • Rate limits: Standard Graph API throttles
  • DR: 2 (own data) / 4 (via Meta Content Library) | SR: 3 | LTV: 3
  • Legal: Standard Meta Platform Terms — strict
  • Dev friendliness: Standard Meta tooling, decent docs
  • Gotchas: No way to search Threads at scale via API. Meta Content Library (March 2026) added Threads but requires academic/researcher approval.

TIER 2: Decentralized / Federated

10. Lemmy

  • URL: https://join-lemmy.org/api / https://github.com/LemmyNet/lemmy
  • Exposes: Posts, comments, communities, votes, federated content via ActivityPub
  • Auth: JWT per-instance
  • Pricing: Free
  • Rate limits: Per-instance configurable
  • DR: 4 | SR: 1 | LTV: 4
  • Legal: AGPL; per-instance
  • Gotchas: Like Mastodon, need to fan out to instances. Federation gives you most content from one instance, but cold-start needs Follow actions.

11. Farcaster (via Neynar + Hubble)

  • URL: https://docs.farcaster.xyz / https://docs.neynar.com
  • Exposes: Casts (posts), reactions, follows, channels, frames, user profiles tied to Ethereum addresses
  • Auth: Neynar API key | Self-hosted Hub = no auth
  • Pricing: Neynar acquired Farcaster Jan 2026 — Neynar API has paid tiers (no free tier per recent docs). Run your own Hubble node = free but minimum hardware.
  • Rate limits: Plan-dependent
  • DR: 4 | SR: 3 (Neynar acquisition consolidates power) | LTV: 3
  • Legal: Open protocol; no token
  • Dev friendliness: Neynar docs excellent, REST + SDKs
  • Gotchas: Post-acquisition uncertainty about open access policy. Run-your-own Hub is the resilient path.

12. Nostr

  • URL: https://github.com/nostr-protocol/nips
  • Exposes: Events (kind 1 = notes, kind 7 = reactions, kind 6 = reposts) via WebSocket relays
  • Auth: None for reads; signed events with secp256k1 keys for writes
  • Pricing: Free (relays free or paid by operator)
  • Rate limits: Per-relay (NIP-11 advertises limits)
  • DR: 3 (signal still low) | SR: 1 | LTV: 3
  • Legal: No central entity. Each relay = own ToS.
  • Dev friendliness: nostr-tools (JS), python-nostr; raw WS protocol
  • Gotchas: Content quality is mostly Bitcoin-maxi; aggregating across relays = engineering lift. No global query — fan out to many relays.

13. Kbin / Mbin / Sublinks / PieFed / Discuit

TIER 3: News Aggregators / Tech Communities

14. Hacker News (Firebase + Algolia)

  • URL: https://github.com/HackerNews/API | https://hn.algolia.com/api
  • Exposes: Stories, comments, users, polls, jobs, real-time updates (Firebase); full-text search + faceted filters (Algolia)
  • Auth: None (anonymous)
  • Pricing: Free
  • Rate limits: Firebase: no published limit (be respectful); Algolia: 10,000 req/hour
  • DR: 4 (high signal-per-byte) | SR: 1 | LTV: 5
  • Legal: No formal ToS issues for read; Y Combinator runs it stably for ~17 years
  • Dev friendliness: Excellent — trivial endpoints, JSON-only
  • Gotchas: Algolia indexing has ~1 min lag. Use Firehose if you need real-time.

15. Stack Exchange (incl. Stack Overflow)

  • URL: https://api.stackexchange.com
  • Exposes: Questions, answers, comments, users, tags, votes, badges across 170+ sites
  • Auth: Optional OAuth 2.0; API key (Stack Apps app) bumps quota
  • Pricing: Free, but data licensing for AI training is being sold separately (deal with OpenAI 2024)
  • Rate limits: Unauth 300/day, with key 10,000/day. Per-IP throttle ~30 req/sec
  • DR: 5 | SR: 2 (AI-licensing aggression is the risk) | LTV: 4
  • Legal: CC BY-SA on content; SO has cracked down on "unauthorized" AI training scraping
  • Dev friendliness: Excellent — generous quota, filters, SDKs
  • Gotchas: Quota_max governs by filter. Bulk archive (archive.org quarterly dumps) is the historical backstop.

16-21. Lobsters, Tildes, Indie Hackers, Product Hunt, Dev.to, Hashnode

See category notes — all are smaller-volume tech-community sources with either lightweight JSON APIs or scrape-friendly endpoints. ProductHunt v2 GraphQL is the most production-ready.

TIER 4: Q&A and Long-Form

22. Quora — no official API, scrape-only via ScrapingBee/Apify/Bright Data. ToS-gray.

23. Substack — RSS official; undocumented JSON endpoints for Notes (C&D risk).

24. Medium — repo archived 2018; only RSS (medium.com/feed/@user) survives. Effectively dead.

TIER 5: Forums (Self-Hostable Platforms)

25. Discourse — REST API, ~50K+ instances incl. Apple, AWS, OpenAI, HuggingFace. Single API surface, very high leverage.

26-30. Vanilla Forums, NodeBB, Flarum, phpBB, XenForo — niche; treat as opt-in customer integrations.

31. Disqus — cross-site comment graph; Zeta Global owned, ad-pivoting.

TIER 6: Imageboards / Anonymous

33. 4chan — official read-only JSON API. Trend/OSINT signal, reputation risk for enterprise.

34. 8kun/Endchan — high reputational risk; not recommended.

TIER 7: Vertical Social

35. Letterboxd — official API rejects LLM use cases. Skip.

36. Goodreads — dead (Dec 2020).

37. Hardcover — Goodreads alternative, GraphQL, free.

38. BookWyrm — federated, ActivityPub.

TIER 8: TikTok and Adjacent

39. TikTok — Research API closed to commercial; commercial route via TikAPI ($49/mo), ScrapeCreators, Apify, Bright Data.

40. Scraping vendors umbrella — Bright Data (legal moat), Apify (6,000 actors), TikAPI, ScrapingBee, Crawlbase, Zyte.

Cross-Cutting Legal Landscape (2024-2026)

  • hiQ v. LinkedIn (settled 2022) — scraping public data not CFAA, but contract claims survive. hiQ paid $500K and destroyed data.
  • X v. Bright Data (dismissed 2024) — public-data scraping without login OK.
  • Meta v. Bright Data (dismissed Jan 2024) — same principle reinforced.
  • LinkedIn v. Proxycurl (filed Jan 2026, settled, Proxycurl shut down July 2026) — fake-account creation was the killer.
  • Reddit v. Anthropic (2025) — pending; precedent-setting on AI training scraping.

Vivly Sequence Recommendation

Phase 0 must-haves (free, low-risk): HN, Stack Exchange, Discourse (covers many big-brand forums in one), Mastodon, Bluesky, Telegram MTProto, Discord (within authorized servers), Slack (within authorized workspaces via Marketplace), Reddit (free tier + scraper fallback).

Phase 1 paid: TikTok via TikAPI/Apify, Quora via ScraperAPI, Indie Hackers via Apify, Substack notes via unofficial (with C&D risk acceptance).

Avoid: Medium (dead), Goodreads (dead), Letterboxd (denies LLM), 8kun-class chans (reputation).

Hedge: Federated protocols (ActivityPub, AT Protocol, Nostr) — lowest stability risk, immune to single-vendor pricing flips. Vivly should disproportionately invest here over closed-platform scraping.