API

Content Intelligence API

Aggregated reports on score, freshness, and linking per project — one API.

The content intelligence API is the catch-all endpoint for everything that measures content quality at the project level: average score, outdated articles, orphan pages, style profile coverage, and the current internal linking state. Instead of one call per sub-system, a single GET gives you a consolidated view — perfect for dashboards and external reporting tools.

Module context: Content Freshness · Internal Linking · Storylines.

Overview endpoint

Method	Endpoint	Description	Credits
GET	`/v1/content-intelligence?project_id={id}`	Aggregated report across all content sub-systems of a project	—
GET	`/v1/content-freshness?project_id={id}`	Project-wide freshness list (which articles are stale)	—

curl "$BASE/content-intelligence?project_id=12" \
  -H "Authorization: Bearer $TOKEN" | jq

Response (truncated):

{
  "data": {
    "project_id": 12,
    "articles": {"total": 142, "avg_score": 78, "below_60": 9},
    "freshness": {"checked": 120, "stale": 14, "last_run_at": "2026-04-29T09:00:00Z"},
    "linking": {"orphans": 3, "suggestions_open": 27},
    "style_profiles": {"active": 2, "coverage_pct": 91}
  }
}

Storylines

Storylines bundle related articles into a topic cluster with shared pillar content.

Method	Endpoint	Description
GET	`/v1/projects/{project}/storylines`	List of all storylines
POST	`/v1/projects/{project}/storylines`	Create new storyline
GET	`/v1/storylines/{id}`	Detail including associated articles
PUT	`/v1/storylines/{id}`	Update
DELETE	`/v1/storylines/{id}`	Delete

Style Profiles

Define tonality, brand voice, and writing rules that generate/optimize will apply.

Method	Endpoint	Description	Credits
GET	`/v1/projects/{project}/style-profiles`	List (includes `source_type`, `source_url`, `pending_review`, `scraped_pages`, `failure_reason`, `reviewed_at`)	—
POST	`/v1/projects/{project}/style-profiles`	Create manually	—
GET	`/v1/style-profiles/{id}`	Detail (same extra fields)	—
PUT	`/v1/style-profiles/{id}`	Update (if `pending_review=true`, implicitly sets it to `false` and records `reviewed_at`)	—
DELETE	`/v1/style-profiles/{id}`	Delete	—
POST	`/v1/projects/{project}/style-profiles/from-url`	Start URL-based style analysis as background job. HTTP 202 → `{ data: { id, status: "pending", source_url, message } }`	25
GET	`/v1/style-profiles/{id}/scrape-status`	Poll job status → `{ status, source_url, scraped_pages, failure_reason, pending_review }`	—
POST	`/v1/style-profiles/{id}/accept`	Accept profile: sets `pending_review=false`, `reviewed_at=now()`. Response: `{ data: <StyleProfile> }`	—
POST	`/v1/style-profiles/{id}/discard`	Discard (soft-delete). Only if `pending_review=true`, else 422. Response: `{ deleted: true }`	—

Knowledge Base

Additional context (PDFs, texts, URLs) that generation pulls in as a source.

Method	Endpoint	Description
GET	`/v1/projects/{project}/knowledge-base`	All entries
POST	`/v1/projects/{project}/knowledge-base`	Add entry
DELETE	`/v1/knowledge-base/{id}`	Remove

Link Lists

Whitelists/blacklists for internal link generation.

Method	Endpoint
GET	`/v1/projects/{project}/link-lists`
POST	`/v1/projects/{project}/link-lists`
DELETE	`/v1/link-lists/{id}`

Orphan Scans

Finds articles with no incoming internal links (SEO problem).

Method	Endpoint	Description
POST	`/v1/projects/{project}/orphan-scans/discover`	Dispatch scan (async)
POST	`/v1/orphan-scans/{id}/start`	Restart
GET	`/v1/orphan-scans/{id}`	Status + summary
GET	`/v1/orphan-scans/{id}/pages`	List of orphan pages
DELETE	`/v1/orphan-scans/{id}`	Delete

SCAN=$(curl -s -X POST "$BASE/projects/12/orphan-scans/discover" \
  -H "Authorization: Bearer $TOKEN" | jq -r .data.id)

# Poll until status == "completed"
curl -s "$BASE/orphan-scans/$SCAN" -H "Authorization: Bearer $TOKEN" \
  | jq '{status, orphans_found}'

curl -s "$BASE/orphan-scans/$SCAN/pages" \
  -H "Authorization: Bearer $TOKEN" | jq '.data[].url'

Notes

The overview endpoint caches internally for 60 s — reports load fast, but values can lag by up to one minute.
freshness.stale only counts articles with a completed freshness check. Articles that have never had a check do not show up in the stale ratio — see Articles API for POST /articles/{id}/freshness/check.
Orphan scans are async, but cheap (no credit cost) and idempotent per project — parallel discovers are deduplicated server-side.
Style profiles only take effect on the next generate/optimize run, not retroactively on existing content.

Letzte Aktualisierung: May 7, 2026