The Hubfluencer API.

Create a Personal Access Token.

curl -X POST https://hubfluencer.com/api/tokens \
  -H "authorization: Bearer $SESSION_TOKEN" \
  -H "content-type: application/json" \
  -d '{"name":"Claude Code","scopes":["video:generate","video:read"]}'

201 Created
{
  "data": {
    "id": 42,
    "name": "Claude Code",
    "scopes": ["video:generate", "video:read"],
    "token": "hf_pat_9f3c…ONCE_ONLY",
    "inserted_at": "2026-06-03T10:00:00Z"
  }
}

List your tokens. A ["*"] scope listing means a full-access token.

200 OK
{
  "data": [
    { "id": 42, "name": "Claude Code",
      "scopes": ["video:generate","video:read"],
      "last_used_at": "2026-06-03T10:05:00Z",
      "inserted_at": "2026-06-03T10:00:00Z" }
  ]
}

Revoke a token. Returns 204 No Content.

Current credit balance. Check it before generating; a short costs 15 credits, and a credit is $0.99 (less in bulk). Quote the dollar cost and get the user's go-ahead before any charging call.

200 OK
{ "data": { "credits": 120 } }

Narration voices. Pass a voice id to generate-voice (editor ads). Shorts have no voice-over.

200 OK
{ "data": [
    { "id": "rachel", "name": "Rachel" },
    { "id": "adam",   "name": "Adam" }
] }

Create a short draft.

curl -X POST https://hubfluencer.com/api/shorts \
  -H "authorization: Bearer $HF" -H "content-type: application/json" \
  -d '{"product_prompt":"a 15s ad for my soy candle brand","language":"en","creative_format":"proof_demo","visual_language":"kinetic_creator"}'

201 Created
{
  "data": {
    "slug": "amber-candle-9x2",
    "stage": "draft",
    "language": "en",
    "latest_render": null,
    "inserted_at": "2026-06-03T10:00:00Z"
  }
}

Generate editable headline, subheadline, and caption beats from the saved short draft (it also fills blank brand_name / closing_claim — extraction and rephrase only, never new claims). No video credits are spent and no render starts. You usually don't need this before generating: generate auto-writes blank copy for free.

curl -X POST https://hubfluencer.com/api/shorts/amber-candle-9x2/text/generate \
  -H "authorization: Bearer $HF"

Render the short. Idempotent per slug. Then poll GET /api/shorts/:slug. Auto-copy: when headline, subheadline, AND text_beats are all blank, the server writes the overlay copy inline first (free — it never touches the daily AI-assist quota) and persists it BEFORE charging, so retries never regenerate it; pass {"skip_auto_text": true} in the body for a deliberately text-free clip. If the copy generator is down → 503 copy_generation_unavailable, nothing charged — just retry. Auto-describe: a product image attached without a description is described inline at generate time (also free — the standalone product-description/generate endpoint stays metered) so the AI segments carry the product grounding; a describer failure just proceeds without one. The prompt (including auto-written copy) is content-compliance screened BEFORE any charge: a prohibited prompt → 422 prompt_rejected (code CONTENT_COMPLIANCE, with a category); screening down → 503 compliance_unavailable (fail-closed) — neither is charged. Out of credits → 402 credits_insufficient with required_credits + available_credits (the legacy required/credits aliases are deprecated).

curl -X POST https://hubfluencer.com/api/shorts/amber-candle-9x2/generate \
  -H "authorization: Bearer $HF" \
  -H "idempotency-key: gen-short:amber-candle-9x2"

Re-render is FREE, forever — credits only ever buy new AI footage. Re-composites the short's CURRENT overlay/style state over the already-generated (pinned) segments + music: edit headline, subheadline, text_beats, cta_text, offer_text, badge_text, star_rating, colors, short_font_family, short_text_position, short_text_animation, accent_color, closing_claim, brand_name, poster (+ poster_includes_lockup), or end_card via PATCH, then call this (no body). Returns the same Short payload as generate — poll GET /api/shorts/:slug. Unchanged composition → the existing render is returned as-is (idempotent). The pinned music bed is reused even when an end-card edit changed the runtime — a shorter bed crossfades into a repeated continuation and fades at the export end instead of failing the render. Footage/music-affecting fields (prompt, product image, creative_format, visual_language, theme, music_vibe, music_instruments, language) need the paid generate instead; changing one returns 422 short_paid_regeneration_required. A FAILED free re-render never breaks the delivered short: stage stays video_ready, latest_render falls back to the latest completed render, and last_free_rerender_failed: true is set (latest_render_free flags any newest free re-render row) — just call rerender again. 409 short_generation_in_progress while any work is active; 422 short_not_ready unless the short has its 2 completed segments, completed music, and a completed render. Never 402.

# edit any free field first (0 credits), then re-render — free, no body
curl -X PATCH https://hubfluencer.com/api/shorts/amber-candle-9x2 \
  -H "authorization: Bearer $HF" -H "content-type: application/json" \
  -d '{"cta_text": "Order today"}'
curl -X POST https://hubfluencer.com/api/shorts/amber-candle-9x2/rerender \
  -H "authorization: Bearer $HF" \
  -H "idempotency-key: rerender-short:amber-candle-9x2:v2"

Poll for status. stage is the signal: draft → rendering → video_ready (read latest_render.video_url) | failed (read failed_stage + error_message). A failed FREE re-render never flips a delivered short to failed — stage stays video_ready with the completed render in latest_render; check last_free_rerender_failed.

200 OK  — poll this until stage is video_ready or failed
{
  "data": {
    "slug": "amber-candle-9x2",
    "stage": "video_ready",
    "failed_stage": null,
    "error_message": null,
    "latest_render": {
      "status": "completed",
      "video_url": "https://…/amber-candle-9x2.mp4?X-Amz-Expires=86400"
    }
  }
}

Cost preflight — returns total and available_credits before you spend.

Edit a draft before rendering: any of the creative fields from POST /api/shorts (product_prompt, headline, subheadline, creative_format, visual_language, music_vibe, the opt-in conversion graphics, …). Present-aware — only the fields you send change. To REMOVE an opt-in graphic, send it empty: offer_text/cta_text/badge_text "", star_rating 0, text_beats []. 0 credits, no render. While the short is generating (processing), edits are locked → 409 short_generation_in_progress; wait for it to finish or fail.

Create a tracking draft. Optional body: language. Returns the slug; upload the source clip via POST /api/editor/:slug/uploads/* (presign or multipart, then confirm) using the returned slug.

List your in-progress tracking projects (the Studio work-in-progress list) — finished ones aren't included. Each entry is the same shape as GET /api/tracking/:slug.

Poll the project. The payload carries source (the latest source upload) and result. source.status is pending → processing → ready (or failed); wait for "ready" before rendering, and treat a non-ready, non-failed source as still uploading/processing. Once rendered, read result.status ("completed") + result.video_url; on a failed render read result.error_message.

Reuse a video already in your asset library as the source clip instead of uploading a new one. Body: asset_id (a ready catalog video asset you own). The asset object is copied into the project and normalized (your library asset is left untouched), so poll GET /api/tracking/:slug until source.status is "ready", then render. 422 asset_not_video / asset_not_ready if the asset isn't a ready video; 409 tracking_render_in_progress while a render is active.

Burn the overlay. Idempotent per slug (send an Idempotency-Key). Optional body: classes (COCO-80 ids to track, e.g. [0]=person, [16]=dog; omit to track all) and color ([r,g,b,a] overlay). An identical re-render returns the existing result for 0 credits. Single-flight: a render started while one is still processing returns 409 tracking_render_in_progress (no extra charge). Out of credits returns 402 credits_insufficient with required_credits + available_credits. Then poll GET /api/tracking/:slug. The source clip is deleted once the render completes (you keep the output); set a new source to render again.

Delete a tracking project and its uploaded clip + result. Deleting mid-render refunds the 1 credit. Requires the video:generate scope (like all write methods).

Create a carousel draft.

Render the carousel. Costs 1 credit per slide ⇒ 3–10 credits (default 5 slides = 5 credits). Send a fresh Idempotency-Key per attempt (a stable per-slug key replays the first response for 24h and blocks an intentional re-generate). The prompt is screened for content-policy compliance first — a violating prompt returns 422 prompt_rejected (code CONTENT_COMPLIANCE, with a category) and is not charged. Rate-limited to 10/min. Then poll GET /api/sliders/:slug until status is completed or failed.

Poll for status until terminal: draft → processing → completed (each slides[].image_url is a downloadable still; read caption + hashtags) | failed (read error_message; credits refunded). Each slides[] entry is { position, headline, body, kicker, status, image_url, background_url }; per-slide status is pending → processing → completed | failed (a not-yet-rendered slide is pending, not draft). image_url is the final composited slide (presigned, 24h TTL); background_url is the raw AI background preview (presigned, 1h TTL).

Edit one slide's on-image text: headline (≤120), body (≤600), kicker (≤40). Re-composites just that slide for free (reuses the AI background — no new image cost). The slider must already be completed (else 409 conflict); it runs async, flipping the slide back to processing — poll GET /api/sliders/:slug until it is completed again. Rate-limited to 20/min.

Restyle a completed carousel: body accepts template, accent_color (^#[0-9a-fA-F]{6}$), text_position (top/middle/bottom), and logo_s3_key — and re-composites EVERY slide for free (reuses the AI backgrounds). The slider must already be completed (else 409). Runs async (every slide → processing); poll GET /api/sliders/:slug until completed. Rate-limited to 20/min. (Confirming a logo via /logo/confirm on an already-completed carousel routes through restyle, so it likewise triggers a full re-composite of every slide, free.)

List your carousels (newest first). GET /api/sliders/:slug/cost returns total + available_credits. PATCH /api/sliders/:slug always edits the copy fields (caption, hashtags); the prompt and the render-shaping fields (language, mode, template, slide_count, aspect_ratio, accent_color, text_position, logo_s3_key) are editable here only while the slider is draft or failed (a failed render reopens them) — once processing/completed they are silently ignored (the prompt is locked post-generation; use restyle / the slide PATCH instead). DELETE /api/sliders/:slug removes a carousel and its images. POST /api/sliders/:slug/logo/presign then /logo/confirm attach an optional brand logo, or POST /api/sliders/:slug/logo/from-asset {asset_id} reuses a ready catalog image you own (subscription-gated, 0 credits).

Create an editor project. Free accounts (0 credits) may hold up to 100 editor projects — a 101st returns 402 editor_factory_limit_reached; delete one or add credits. Accounts with credits are uncapped. Send an Idempotency-Key header so a retry doesn't fork a second draft.

curl -X POST https://hubfluencer.com/api/editor \
  -H "authorization: Bearer $HF" -H "content-type: application/json" \
  -d '{"language":"en","product_prompt":"a cinematic ad for my candle brand","export_aspect_ratio":"9:16"}'
# then start autopilot:
curl -X POST https://hubfluencer.com/api/editor/$SLUG/autopilot \
  -H "authorization: Bearer $HF" -H "idempotency-key: autopilot:$SLUG"

Run the full pipeline. Send an Idempotency-Key. Preflight the cost with GET /api/editor/:slug/autopilot/cost. Accepts an OPTIONAL JSON body: language (one of the supported project codes), product_subject (3–300 chars when non-blank), product_prompt (10–5000 chars when non-blank), and max_credits (non-negative integer). A structured scenario brief, product_prompt ≥10, or product_subject ≥3 satisfies the brief requirement. The cap is persisted for this run and atomically enforced again at every scene, voice, and music charge; if it would be exceeded, no further charge starts. Only non-blank text overrides are applied before the chain starts.

curl -X POST https://hubfluencer.com/api/editor/$SLUG/autopilot \
  -H "authorization: Bearer $HF" -H "content-type: application/json" \
  -H "idempotency-key: autopilot:$SLUG" \
  -d '{"language":"fr","product_subject":"Bordeaux red wine, dark green bottle with a gold label","product_prompt":"a cinematic 15s ad for my wine","max_credits":28}'

200 OK  — poll this; autopilot embeds the finished MP4 in latest_render
{
  "data": {
    "slug": "candle-story-7k1",
    "autopilot_status": "completed",
    "autopilot_error_message": null,
    "scenario_prompt": "Warm, handcrafted… ",
    "segments_count": 5,
    "segments": [
      { "id": 81, "position": 1, "prompt": "Macro of wax pouring…", "status": "completed" },
      { "id": 82, "position": 2, "prompt": "Hands trimming the wick…", "status": "completed" }
    ],
    "narration_script": "Made by hand, poured in small batches…",
    "narration_status": "completed",
    "music": { "status": "completed" },
    "latest_render": {
      "status": "completed",
      "video_url": "https://…/candle-story-7k1.mp4?X-Amz-Expires=86400"
    }
  }
}

Stop a running autopilot chain (autopilot_status → cancelled). Already-charged scenes stay; nothing new is generated. No-op if it already finished.

Get a presigned PUT URL for a video upload.

Returns 422 upload_quota_exceeded (with a quota object: limit/used/reserved/remaining bytes) if the upload would push your total live upload storage over the limit. Delete unused uploads to free space.

200 OK
{
  "data": {
    "upload_id": 510,
    "presigned_url": "https://r2…/editor/$SLUG/uploads/uuid.mp4?X-Amz-Signature=…",
    "s3_key": "editor/$SLUG/uploads/uuid.mp4",
    "expires_in_seconds": 3600
  }
}

Confirm the PUT. The server validates the object (Content-Type and size must match the presign) and queues processing.

List uploads with their processing status. Poll until status is "ready" before placing the clip.

200 OK  — status flows pending → processing → ready | failed
{
  "data": [
    { "id": 510, "filename": "clip.mp4", "mime_type": "video/mp4",
      "status": "ready", "duration_seconds": 6.2, "width": 1080, "height": 1920,
      "error_message": null, "attached_segment_ids": [83] }
  ]
}

Append a ready upload to the timeline as a finished scene. Body: upload_id. Rejects editor_upload_not_ready, batch_generation_active, autopilot_active (while Autopilot is running), editor_segment_limit (max 20 scenes).

Set one scene's video from an existing asset. Body: exactly one of upload_id or source_segment_id. Reuse a clip across scenes, or swap a generated scene for your footage.

The 200 data is the updated segment plus an impact object describing downstream continuity fallout — replacing a scene changes the frame the next continuity-enabled AI scene inherits. Keys: requires_regeneration (bool), stale_video_segment_ids / stale_video_positions (following completed AI scenes now stale), affected_segment_ids / affected_positions (the whole downstream chain), anchor_segment_id / anchor_position, stop_reason / stop_position, and action ("segment_asset_replace"). When requires_regeneration is true, re-generate the listed scenes before rendering — render rejects a stale timeline.

Every render version with status and a presigned video_url. Use it to recover a finished URL or find a failed render to retry.

200 OK
{
  "data": [
    { "id": 9, "version": 2, "status": "completed",
      "video_url": "https://…/render.mp4?X-Amz-Expires=86400",
      "duration_seconds": 18.0, "error_message": null,
      "inserted_at": "2026-06-03T10:20:00Z" },
    { "id": 8, "version": 1, "status": "failed",
      "video_url": null, "error_message": "forge_timeout" }
  ]
}

Re-run a failed render from its saved snapshot. Only failed renders are retryable (editor_render_not_retryable otherwise).

Fetch a public product page (HTML only) and return parsed facts: name, description, benefits, price, brand, availability, plus candidate product image URLs and a site/brand logo URL. URLs only — images are NOT downloaded here. Body: {url} (absolute http(s), ≤2048 chars). 422 unsupported_content_type for a non-HTML page, too_large for an oversized one.

Fetch an image at a public URL (JPEG/PNG, verified by magic-number sniffing — not just the Content-Type — and capped to a small preview) and return an inline data URL for review. Writes nothing, stores no asset. Body: {image_url}. 422 for an unsafe or non-image URL.

Fetch an image at a public URL (JPEG/PNG magic-number verified, size-capped at the catalog image limit) and store it as a ready CatalogAsset. Body: {image_url, product_profile_id?, link_as?}. With product_profile_id it links the asset to that profile as link_as "primary_image" (default) or "logo". 404 for an unknown/foreign profile; 422 for an unsafe or non-image URL.

List / read your product profiles. GET /api/product-profiles/:id returns one. Pass an id to plan / hook-variations / campaign-pack create so the campaign is planned from a saved source of truth.

Create a product fact sheet. Body: {name (required), description?, benefits?[], offer?, proof_points?[], audience?, cta?, banned_claims?[], brand_tone?, source_url?}. PATCH /api/product-profiles/:id patches any of these (0 credits); DELETE /api/product-profiles/:id removes it (204). Attach imagery via product-sources/import-image. 404 for an unknown/foreign id.

List / read your brand profiles (GET /api/brand-profiles/:id for one). Pass an id as brand_profile_id to POST /api/shorts / hook-variations / campaign-pack create / series so drafts inherit the brand + logo.

Create a brand identity. Body: {name (required), logo_s3_key?, primary_color?, secondary_color?, font_family?, tone_words?[], default_cta?, default_audience?, banned_claims?[], preferred_visual_language?, preferred_creative_format?}. Colors are hex (#09EFBE); preferred_visual_language / preferred_creative_format take the same value lists as Shorts/editor. PATCH /api/brand-profiles/:id patches any field; DELETE removes it (204).

Attach one of your catalog assets to a brand profile. Body: {catalog_asset_id, role} where role is logo | product_image | other. DELETE /api/brand-profiles/:id/assets/:asset_id removes a link — :asset_id is the LINK id (the profile's assets[].id), not the catalog_asset_id. 422 if the catalog asset isn't yours; 404 for an unknown/foreign profile.

Turn product facts into a campaign PLAN: hooks, recommended formats, a positioning line, and a per-item credit ESTIMATE. Persists nothing. Body: {product_profile_id | product{name,...}, goal?, platform?, language?}. Spends 0 credits but consumes 1 free AI assist (429 ai_assist_quota_exceeded when the daily quota is spent). Rate-limited to 30/min.

Turn product facts + an optional brand_profile_id into 3–6 distinct hook variations, each with a hook, angle, suggested output format (editor_ad/short/carousel), a draft script, a caption, hashtags, and a per-variation credit estimate. Persists nothing. Body: {product_profile_id | product{...}, brand_profile_id?, count? 3..6, goal?, platform?, language?}. 0 credits, 1 free AI assist. Rate-limited to 30/min.

Persist a pack plus any inline hook/caption TEXT items. editor_ad/short/carousel items are added via /variations (below), not here. Body: {title?, product_profile_id?, brand_profile_id?, brief?, source_type?, source_url?, language?, items?:[{kind:"hook"|"caption", ...}]}. Link a product/brand profile so materialized drafts inherit the image + logo. GET /api/campaign-packs lists your packs; GET /api/campaign-packs/:id returns one with its items.

Persist selected hook variations as items under the pack and, by default (materialize:true), turn each into a DRAFT. Body: {variations:[{suggested_format:"editor_ad"|"short"|"carousel", hook?, angle?, draft_script?, caption?, hashtags?, credit_estimate?}], materialize?}. At most 12 variations per call; rate-limited to 15/min. Per-variation best-effort: one that can't be materialized stays a "planned" item with a recorded reason while the rest still materialize — the response summary lists created/materialized/planned + skipped_reasons.

Materialize one planned editor_ad/short/carousel item into a DRAFT and link it back to the item. Idempotent — re-calling a just-materialized item returns the existing draft (no duplicate). Rate-limited to 60/min. 422 for a hook/caption item, an already-materialized item, or when you hit a draft limit. Each item carries its status (planned → draft → generating → completed | failed), the draft slug + deep-link route (editor_ad/short → the video project, carousel → the slider), and a credit estimate. Generate each draft EXPLICITLY next: POST /api/editor/:slug/autopilot for videos, POST /api/sliders/:slug/generate for carousels.

List / read your series (GET /api/series/:id for one). GET /api/series/:id/dashboard returns the series plus its episodes grouped into lanes — upcoming (ideas), drafted, published — with per-item performance notes.

Create a show template. Body: {name (required), template (required), brand_profile_id?, product_profile_id?, cadence?, tone?, default_format?, status?}. default_format is editor_ad/short/carousel; status toggles active/paused/archived (a paused series stops surfacing new ideas). A supplied brand/product profile must be your own (422 invalid_brand_profile / invalid_product_profile). PATCH /api/series/:id patches any field; DELETE removes it (204).

Draft 3–5 upcoming episode ideas (status "idea") for the series, grounded on its product profile and steered by its brand/tone/cadence. Body: {count? 3..5, language?}. Persists them as episodes. 0 credits but consumes 1 free AI assist (429 when the quota is spent). Rate-limited to 30/min. GET /api/series/:id/episodes lists episodes (optional ?status filter).

Turn an "idea" episode into an editable 0-credit DRAFT under the series' rolling pack, inheriting the series' brand. Idempotent — re-calling a just-materialized episode returns the existing draft. Rate-limited to 60/min. 422 for an already-materialized / non-idea episode or a draft limit. The episode carries its draft slug + route; generate it explicitly next.

Record that you posted a DRAFTED episode manually (drafted → published, stamps posted_at) — there is no platform posting API. Idempotent: re-marking a published episode returns 200 with the original posted_at. 422 invalid_status for an idea/archived episode (materialize it first); 404 for an unknown/foreign series or episode.

Record one performance snapshot. Body: {platform, views?, likes?, comments?, shares?, clicks?, conversions?, revenue?, watch_time_seconds?, source?, captured_at?, metadata?}. Append-only — many snapshots per item (captured_at, defaulting to now, distinguishes them), NOT idempotency-keyed, so don't blindly retry. Rate-limited to 30/min. GET /api/campaign-pack-items/:item_id/performance lists them (newest first) with a small summary. 404 for a foreign/unknown item.

Insert-or-update the SINGLE creative-attributes row the recommendation loop correlates with performance. Body: {hook_type?, creative_format?, visual_language?, product_category?, cta?, length?, caption_style?, metadata?}. Atomic upsert — calling it again updates the same row. GET /api/campaign-pack-items/:item_id/attributes returns the row or null. 404 for a foreign/unknown item.

Aggregate your items that have BOTH attributes AND ≥1 snapshot into a cautious recommendation: the top hook_type / creative_format / visual_language by average views, your best items, and an HONEST confidence ("none" / "low" / "moderate", a function of sample size) plus a non-causal note. It NEVER claims causation, ROAS, or attribution — these are correlations in your own data, a hint to test. Present the note + confidence as-is. No LLM, no quota, no credits.

Generate 3 follow-up hook variations from a winning item, threading its angle/format/hook while keeping its brand + product profile. 0 credits but consumes 1 free AI assist (429 when spent). Rate-limited to 30/min. 404 for a foreign/unknown item; 422 when the item has no usable product context. Feed a chosen variation into /variations to draft it.

Create a review link for exactly one of your own COMPLETED outputs. Body: {video_result_id | slider_id (exactly one), campaign_pack_id?, expires_in_hours?, download_enabled?, version_label?}. expires_in_hours defaults to 168 (7 days) and is clamped to a supported range. 422 for a not-completed/foreign output or a foreign campaign pack. Returns {id, token, review_url, expires_at, ...} — the token appears ONCE.

List your review links (newest first) with status (pending → viewed → approved | changes_requested | revoked) and a visible comment_count. Never returns the token. GET /api/review-links/:id returns one. DELETE /api/review-links/:id revokes a link (status → "revoked"; the URL stops working immediately).

List your catalog (newest first), each with media_type, mime, size, description, status, and a presigned asset_url; meta.quota carries your storage quota. Optional ?limit / ?offset (offset = page * limit). GET /api/assets/quota returns just the quota (limit / used / pending / reserved / remaining bytes).

Presign an upload. Body: {filename, mime_type, size_bytes} → { data: {asset_id, presigned_url, s3_key} }. PUT the raw bytes to the presigned_url with header Content-Type: <the presigned mime>, then POST /api/assets/:id/confirm {duration_seconds?, width?, height?, description?} — confirm HEADs the object and validates size/type/duration (422 on mismatch). Images ≤ 20 MiB; videos ≤ 500 MB and REQUIRE duration_seconds ≤ 60. Rate-limited to 30/min.

List your scheduled posts, ordered by scheduled_at ascending. Optional query params: ?from= / ?to= (ISO-8601 UTC bounds on scheduled_at) and ?status= (scheduled | reminded | posted | publishing | published | failed | canceled).

Fetch one of your scheduled posts. 404 for unknown/foreign ids.

Schedule one of your COMPLETED renders for a future UTC instant (scheduled_at) on a target platform. WITHOUT a social account id this is a reminder-only schedule — an agent token (video:generate) can create it. WITH an account id (tiktok_account_id for "tiktok", instagram_account_id for "instagram", one of your own accounts) it is an AUTO-PUBLISH that additionally requires account:admin (403 insufficient_scope otherwise) AND that platform's posting feature (403 feature_disabled). Optionally grouped under a campaign pack item. Persists the schedule only; spends NO credits. At most 500 pending posts per user (422 scheduled_post_limit_reached beyond).

Reschedule a "scheduled" or "reminded" post to a new FUTURE scheduled_at (a reminded post is re-armed to "scheduled"). video:generate manages a reminder-only post; a post that carries a social account (or an update attaching one) additionally requires account:admin (403 insufficient_scope). 422 not_reschedulable / scheduled_in_past / invalid_scheduled_at; 404 for unknown/foreign ids.

Cancel a "scheduled", "reminded", or "failed" post (status -> "canceled"). video:generate cancels a reminder-only post; canceling a post that carries a social account additionally requires account:admin (403 insufficient_scope). 422 not_cancelable for an already posted/publishing/published/canceled post; 404 for unknown/foreign ids.

Confirm you posted this content manually: transitions a "scheduled" or "reminded" post to "posted" and stamps posted_at. Atomic — a post in any other status (including an already-posted one, so a second call) returns 422 not_markable. 404 for unknown/foreign ids. Touches no platform API; spends NO credits. Lets an agent close the reminder loop after posting.

The current quota.

200 OK
{
  "data": {
    "used": 4, "limit": 20, "bonus": 0, "remaining": 16,
    "resets_at": "2026-06-04T00:00:00Z",
    "unlock_cost": 1, "unlock_batch_size": 10
  }
}

Spend 1 credit for +10 assists. It's a repeatable purchase — do NOT send a stable Idempotency-Key (a reused key replays the first unlock for 24h, silently skipping later purchases).