Skip to content

docs: comprehensive update to match Late API as of March 2026#3

Open
s7an-it wants to merge 2 commits intogetlate-dev:mainfrom
s7an-it:fix/update-docs-march-2026
Open

docs: comprehensive update to match Late API as of March 2026#3
s7an-it wants to merge 2 commits intogetlate-dev:mainfrom
s7an-it:fix/update-docs-march-2026

Conversation

@s7an-it
Copy link

@s7an-it s7an-it commented Mar 5, 2026

Summary

Comprehensive update to match the Late API as of March 2026, plus skill quality improvements based on benchmarking against common integration scenarios.

Accuracy issues fixed

  • GET /v1/connect/{platform} (was documented/used as POST /profiles/{id}/connect) — the single most common mistake
  • scheduledFor field name (was scheduledAt in examples)
  • customContent for platform-specific text (was sometimes shown as overrides)
  • publicUrl from presign response (not uploadUrl) must be used in post body
  • redirectUrl is required in connect — added to Quick Start example
  • Docs URL: getlate.dev/docsdocs.getlate.dev

New content (March 2026 changelog)

  • rules/profiles.md — was missing entirely; documents Profile CRUD endpoints
  • rules/analytics.md — 5 new endpoints: daily-metrics, best-time-to-post, content-decay, posting-frequency, post-timeline
  • rules/tools.md — 4 new validation endpoints: post-length, post, media, subreddit
  • rules/errors.md — HTTP 409 DUPLICATE_CONTENT (Jan 2026); platform error fields errorMessage/errorCategory/errorSource (Feb 2026); partial post status
  • rules/integration-guide.md — new end-to-end walkthrough: profile → OAuth connect → media upload → post → webhook

Skill improvements

  • Common Gotchas table: 9 mistakes Claude makes without reference docs (validated by benchmark)
  • Response schemas for POST /v1/posts, GET /v1/connect/{platform}, POST /v1/media/presign
  • Post status values documented (scheduled/published/failed/partial)
  • Pricing context table with SaaS guidance
  • Improved skill description for better triggering

Benchmark results

Tested against 3 realistic integration scenarios (Reddit OAuth flow, multi-platform scheduling with media, debug 405 error). Without this skill: 8.3% pass rate (1/16 assertions). With this skill: 100% (16/16).

thinkmakeshipiot and others added 2 commits March 5, 2026 03:20
- Add profiles.md: CRUD endpoints for profile management (POST, GET, PATCH, DELETE)
- Update connect.md: fix redirectUrl/authUrl field names, add headless OAuth mode, Reddit subreddit selection flow, Telegram access code flow
- Update media.md: fix fileName/fileType/publicUrl fields, add customMedia per-platform, auto-compression docs, correct 5GB upload limit
- Update platforms.md: add complete Reddit platformSpecificData section (subreddit, title, flairId, url, forceSelf)
- Update accounts.md: add reddit-flairs endpoint, Reddit-specific curl examples
- Update posts.md: add unpublish endpoint, customContent overrides, tags, accountIds shorthand
- Update analytics.md: add 5 new endpoints from Feb-Mar 2026 (daily-metrics, best-time-to-post, content-decay, posting-frequency, post-timeline)
- Update tools.md: add 4 validation endpoints from Feb 2026 (post-length, post dry-run, media, subreddit)
- Update errors.md: add 409 duplicate content detection (Jan 2026), platform target error fields with errorCategory/errorSource (Feb 2026)
- Update SKILL.md: reference profiles.md, update descriptions for all rule files

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…xt, fix docs URL

- Fix docs URL: getlate.dev/docs → docs.getlate.dev
- Improve skill description for better triggering
- Add Common Gotchas table (9 most-common mistakes without reference docs)
- Add end-to-end integration-guide.md: profile → connect → media → post → webhook
- Add response schemas for POST /v1/posts, GET /v1/connect/{platform}, POST /v1/media/presign
- Add post status values (scheduled/published/failed/partial)
- Add pricing context table with SaaS guidance (Accelerate minimum)
- Add redirectUrl requirement to Quick Start connect example

Benchmark: skill achieves 100% accuracy (16/16 assertions) vs 8.3% baseline
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants