Simplified brings your marketing workspace into ChatGPT. Generate AI images and video, create speech from text, draft and publish social posts across 13 connected channels, manage brand kits, projects and tasks, plan content calendars and campaigns, and analyze social performance — all without leaving the conversation.
object_pk to the resource UUID and content_type to the resource family (currently task). For threaded replies, set parent to the parent comment ID.brand — name, description, website URL - social_links — array of {type, url} objects - content_pillars — 1–5 grounded themes, merged by normalized name - style — full visual-identity JSON (colors, typography, logos, imagery, logo_lockup, on_image_copy, ai_guardrails, prose, inferred_fields) The body shape is identical to the GET response (without id). Empty body {} is a valid no-op. Server-controlled fields (id, version) in the body are ignored. Runs synchronously and returns 200 with the result. Calling this endpoint multiple times is safe — sections sent overwrite their own fields; omitted sections are left unchanged. Content pillars omitted from a supplied list are preserved. Do NOT call `getTaskResult` after this — there is no async task; the operation is complete by the time the response returns. Path Parameters: - brand_id: Brand kit UUIDurl — a remote file URL the server downloads, OR - path — a local file path; the toolkit uploads it via a signed URL and registers the asset (only works when running smp / smp serve locally — a remote MCP host can't read your filesystem). Processing is async (S3 upload + thumbnail generation via Celery). The returned asset UUID is usable immediately — e.g. in createProjectItem data.assets[] — but poll getAsset until status=4 (DONE) before relying on the stored URLs. Common use: persist a transient AI-generated image (from generateImage) so it can be saved to a project, published, or opened in the editor.title is required — this makes an empty V2 kit. Populate it afterwards with buildBrandKit (canonical brand/style document) and/or context documents (createContextDocument). Typical onboarding flow: 1. createBrandKit with the brand name as title → returns the kit id (UUID) 2. buildBrandKit with the canonical BrandKitDocument to fill brand + style 3. getBrandKit?expand=extra to read it backdocument_id to link an existing KnowledgeDoc. 2. Inline creation: Provide doc_type and name (plus optional description, data, content) to create a new KnowledgeDoc and link it automatically. Singleton constraint: For predefined types (brand_voice, icps, usps, etc.), only one document per type per brandkit is allowed. Attempting to create a duplicate will return a 400 error. The canonical_key is auto-set from doc_type if it matches a predefined type. Path Parameters: - brand_id: Brand kit UUID/ai-writer/document/{space_or_organization}/{id} in the Simplified web app. To convert markdown content into the required Quill Delta shape, call markdown_to_quill_delta from the writers toolkit before passing to this endpoint.resourcetype. title is the common field; type-specific fields go under data (varies by resourcetype — inspect a getProject response for the target type to see its shape). Returns the new project's UUID, used for createProjectItem, listProjectItems, etc. Path Parameters: - resourcetype: Project type selector (the model name). The API is polymorphic — the same URL structure serves different project types based on this value. Valid values: - Project — generic projects (the default for most content). - AdCreativeProject — ad-creative projects (specialized fields). Use the same value on create, get, list, and delete for a given project.is_deleted=true. The project and its items stop appearing in listProjects/getProject but are not purged from the database, so this is recoverable server-side. resourcetype must match the project's type (Project or AdCreativeProject) — the same value used to create it. Path Parameters: - resourcetype: Project type selector - id: Project ID (UUID or integer depending on type)is_deleted=true. The item stops appearing in listProjectItems/getProjectItem but is not purged, so it is recoverable server-side. Requires the parent project ID and matching resourcetype. To remove the whole project, use deleteProject. Path Parameters: - resourcetype: Project type selector - parent_lookup_project_id: Parent project ID - id: Item ID (UUID)voice_id (get voices from listVoices). storage controls whether the audio is saved. Set use_ssml: true ONLY when text contains SSML markup like <speak> / <break> / <emphasis>. For plain prose leave it false — providers handle punctuation pacing well and SSML support varies. Storage modes: - default — returns the audio URL only (can expire) - asset — also saves it to your asset library for reuse Path Parameters: - voice_id: Voice UUID from listVoices voices[].id{model, parameters} (capability is optional). It blocks and returns the final result — do not call getTaskResult. Use the field names below as-is for every model; the server adapts the payload to the model you chose. If something does not fit that model (a ratio it lacks, too many references for the mode), the error says exactly what to change. storage controls where the result is kept: - "asset" — saved to your asset library for reuse (recommended) - "transient" — temporary, not saved - "default" — kept in the image gallery{model, parameters} (capability is optional — derived from which inputs you send). It blocks and returns the final result — do not poll. Use the input slot names below as-is for every model; the server adapts the payload to the model you chose. If something does not fit that model (a ratio, duration or resolution it lacks, too many images for the mode), the error says exactly what to change. getModelFields (type=video) is optional — call it only when you want the model's duration/resolution options before choosing. Why task_id is misleading here The endpoint also returns a task_id, but it tracks the submission Celery task that hands the request off to FAL/Kie — it flips to SUCCESS almost immediately, well before the actual video is rendered. Do not call `getTaskResult` on this `task_id` — it tells you nothing useful about generation completion. Always poll getVideoVariation instead. The apikit middleware strips task_id from this response automatically to avoid confusion. Storage modes - default — persists into the AiImageArt gallery (recommended). - transient — generated without surfacing in the gallery; useful for one-shot agent runs. - asset — saves the rendered video as a standalone reusable TldrAsset (no gallery entry).createAsset from a URL, or a signed upload) — the asset UUID is returned immediately but the file download / S3 upload / thumbnail generation completes in the background. Key response fields: - status — processing state: 0 NOT_STARTED, 1 IN_PROGRESS, 2 PROCESSING_FAILED, 3 THUMBNAIL_FAILED, 4 DONE, 10 NOT_PROCESSED. Poll until 4 before relying on the URLs. - file_url / thumbnail — signed URLs to the stored file (expire, ~12h; re-fetch to refresh). - asset_url — non-signed stored URL. - asset_type — image / video / audio / pdf / etc. - payload — extracted media metadata (width, height, duration, format, ...). Path Parameters: - id: Asset UUID (e.g. the id returned by createAsset).{id, title} only. Use ?expand=extra to include the full brand envelope (brand, social_links, style). Use ?expand=website for the resolved company website URL. Combine with ?expand=extra,website. Sparse fieldsets via ?fields= and ?omit= are honored on the top-level keys. Path Parameters: - brand_id: Brand kit UUIDdocument_link_id may be EITHER the junction (BrandKitContextDocument) UUID returned by listContextDocuments, OR the underlying KnowledgeDoc UUID — the server resolves either form. Use this when you already have a document ID and want its content. To find a document by its canonical type (e.g. brand_voice) instead, use getContextDocumentByType. To list all documents on a kit, use listContextDocuments. Path Parameters: - brand_id: Brand kit UUID - document_link_id: Either the context-document link UUID (BrandKitContextDocument ID) OR the underlying KnowledgeDoc UUID — the endpoint resolves both.content_pillars, the backend reads the Brand Kit's existing structured ContentPillar records, renders cached Markdown on demand, and returns {doc_type: "content_pillars", content: "..."}. It does not create or synchronize a KnowledgeDoc. If no structured pillars exist, the endpoint returns 404 even when a legacy active context document exists. Path Parameters: - brand_id: Brand kit UUID - context_type: Canonical type key to retrievecredits object: - credits.remaining — credits still available to spend - credits.used — credits consumed in the current period - credits.total — total credits allotted for the period - credits.trial — extra trial allowance (only present when on trial)google.gemini-3.1-flash-image or flux.flux-realism.parameters object of generateImage from what this returns rather than reusing settings from a different model.generateVideo. Not needed for generateImage, which takes one fixed contract for every model. A single unified endpoint serves both image and video engines, selected by the required type query param. Response modes (driven by which query params you pass): - type only → {models: [...]} — list every registered engine of that type, its capabilities, and metadata (incl. credits_per_image / estimated_time). Use this to pick a model_id. - type + model_id → single-model detail with capabilities map. - type + model_id + capability → the full FieldDescriptor schema for that (model, capability) pair (labels, field_type, required flags, enum_values, defaults). This is the contract that generateImage / generateVideo's parameters must match. Use this before generating — it eliminates 400 errors on invalid / missing parameter keys.task_id. Status values: PENDING, STARTED, SUCCESS, FAILURE, RETRY, REVOKED. Note: for most async tools (generateImage, media/ operations) the toolkit middleware already polls this for you and returns the final result — you rarely need to call it directly. Use it manually only when a long job (e.g. video generation) exceeds the auto-poll window and you got a `task_id` back. Path Parameters: - task_id*: Celery task IDveo-3.1, sora-2, kling-v2.5-turbo-pro.capability is required, because a video model's fields differ substantially between capabilities — a first_last_frame call takes asset ids where a prompt call takes text.generateVideo. Read job_status: - Non-terminal: CREATED, PENDING, PROCESSING, RENDERING, UPDATED — keep polling. 10s is a reasonable interval. - Terminal: DONE (success — payload.output and asset_references are now populated with the rendered video URL) or FAILED (error — inspect payload.errors). Generation is webhook-driven (FAL / Kie call back into djapp when rendering finishes), so the row updates asynchronously — there is no task_id to short-circuit polling. Path Parameters: - art_id: AiImageArt UUID returned as id from generateVideo. - variation_id: AIArtVariation UUID returned as art_variation_id from generateVideo.voice_id — one the user picked in a voice picker, or a saved preference — back to its attributes without re-listing the whole catalog: gender (0=Male, 1=Female, 3=Neutral), voice_name, is_premium, description, and supports_instructions (whether the voice honors the instructions hint on generateAudio). The response is a single voice object — the same shape as one entry of listVoices's voices[]. Resolution is workspace-scoped: a voice owned by another workspace (and not globally published) returns 404. Path Parameters: - voice_id: Voice UUID (the id from a listVoices voices[] entry).Organization header. Response includes: - user — id, first_name, last_name, email of the requesting user - workspace — id, name, slug, is_active, and settings (timezone, language, start_of_week) - teamspaces — active spaces visible in workspace context (id, name, slug, is_active) Do not use this response as the membership authority for space_id. Call listTeamspaces to resolve IDs the current user can scope into.search to list them, newest first; pass it to narrow by asset name or tag. search is text matching, not visual or semantic. A search term includes matching assets stored inside folders; it does not search folder names or return every asset in a named folder. asset_type narrows either form, so listing only videos needs no search term. Results contain asset UUIDs (id), asset_name, asset_type, status, and thumbnail metadata. Use the UUIDs directly as social post media. Call getAsset for full asset details and fresh stored URLs. Follow pagination by incrementing page with the same search and filters when the response indicates another page. Do not stop after the first page when the user asks for all matching assets.getBrandKit, buildBrandKit, and the context-document operations. Use ?search=<title> to filter by name when the user refers to a kit by title. Paginated — pass page_size to widen results.object_pk (resource UUID) and content_type (resource family, e.g. task).canonical_key, searching by document name or type, and ordering by created/modified timestamps. Path Parameters: - brand_id: Brand kit UUIDgenerateImage, with the capabilities each one supports. Public — no credential required. Call this before generating when you do not already know a model id. Passing a setting a model does not support is silently ignored rather than rejected, so discovering the model first is how you avoid an image that comes back subtly wrong (an aspect ratio that never applied, for example).resourcetype in the workspace, paginated and newest-first. Returns project objects with their UUIDs — needed for getProject, listProjectItems, and deleteProject. Filter with ?search=<title>. Pass the same resourcetype you created the projects with. Path Parameters: - resourcetype: Project type selector (the model name). The API is polymorphic — the same URL structure serves different project types based on this value. Valid values: - Project — generic projects (the default for most content). - AdCreativeProject — ad-creative projects (specialized fields). Use the same value on create, get, list, and delete for a given project.id (integer) is the value you pass as the Space header (or --space / space_id) to narrow a request to that teamspace. The token stays bound to its one workspace — this does not switch workspaces, it only lists the sub-spaces you can scope into. Supports ?search=<name> to filter by teamspace name and ?expand=settings. Each item: {id (int), name, slug}.generateVideo, with supported capabilities, durations, resolutions and aspect ratios. Public — no credential required. Video models vary far more than image models — duration and resolution options are per-model — so call this before generating.language_code is the ONLY required parameter — use an exact BCP-47 locale (en-US, en-GB, fr-FR, hi-IN, ja-JP). Combine with is_premium=true for a short, high-quality shortlist. The optional filters (gender, provider, engine, is_premium, language_name, search) all default to an empty string "", which means "no filter — return everything". Leave them at "" unless the user explicitly asked to filter. Do NOT substitute 0/false for "no filter": 0 is a REAL value (gender 0 = Male, provider 0 = Google, engine 0 = Neural), so sending it silently narrows the list. The empty-string default is the ONLY correct "unset" value. Response: - voices[]: id (uuid), voice_name, language_code, avatar, voice_sample, gender (0=Male, 1=Female, 3=Neutral), is_premium, description, supports_instructions (bool — true if the voice honors the instructions style/pace hint on generateAudio). Use id as voice_id in generateAudio. To re-resolve a single known voice_id (e.g. from a saved preference) without re-listing, use getVoice. - languages[]: distinct {language_code, language_name} pairs. To resolve a known voice_name (e.g. from an avatar roster) to its id, pass voice_name + language_code — that pair is unique, and the id differs per locale, so the name must be re-resolved whenever the language changes. Always check the entry you use actually has that voice_name rather than trusting voices[0] — see the voice_name parameter.signAssetUpload. Carry every registration field from that response unchanged and use the original filename for asset_name. Returns the permanent asset UUID used by social media.signed URL using the declared MIME type and no Simplified Authorization header. After the PUT succeeds, call registerAsset with the returned metadata.data, first read the project and send the complete merged data object so unrelated state is retained. Path Parameters: - resourcetype: Project type selector - id: Project ID (UUID or integer depending on type)data, first read the item and send the complete merged data. Path Parameters: - resourcetype: Project type selector - parent_lookup_project_id: Parent project ID - id: Item ID (UUID)image_url and the target output_format. This changes the file format ONLY — for pixel-level edits (background removal, upscale, inpaint) use the other image tools in this module. Async — returns a task_id (auto-polled).output_format (e.g. mp4, mov, webm, gif). Changes format only — not resolution or content. Async — returns a task_id (auto-polled).getTranscription returns job_status="DONE"; otherwise the file body will be empty. Response is the raw subtitle text. Path Parameters: - id: Transcription UUID.prompt. Requires a mask (mask_url or mask_base64) marking WHERE to fill. Use this when you have a specific region to replace; for removing an object by description (no mask) use magicInpaint; to extend the image past its edges use imageOutpainting. Async — returns a task_id (auto-polled).id returned by transcribeVideo. Read job_status: - Non-terminal (CREATED, PENDING, PROCESSING, RENDERING, UPDATED): keep polling. A 10s interval is reasonable. - Terminal: DONE (success — payload.phrases is now populated with word-level timing) or FAILED (error — stop polling). Once job_status="DONE" you can either read payload.phrases directly, or call downloadTranscriptionFile for a VTT/SRT file. Path Parameters: - id: Transcription UUID returned from transcribeVideo.prompt. Requires mask_url marking the new area. Different from generativeFill (fills a region inside the image) and magicInpaint (edits existing content). Async — returns a task_id (auto-polled).prompt, with no mask needed (the model locates the target). Use for "remove the person" / "replace the car with a bike". With an explicit mask region use generativeFill; to extend beyond the borders use imageOutpainting. Async — returns a task_id (auto-polled).video_urls (needs 2+ URLs). The result is a single clip. Async — returns a task_id (auto-polled).replace_type to color (needs replace_color), image (needs replace_image), or transparent. To blur rather than replace use blurBackground; to cut the subject out entirely use removeBackground. Async — returns a task_id (auto-polled).task_id; poll getTaskResult (the toolkit auto-polls) for the output URL. Audio is reversed along with the video. For trimming or speed changes use speedupVideo; this only reverses direction.prompt (and negative_prompt). Pass the scribble as image_url or base64 image. Generates a NEW image from line art; to edit an existing photo use magicInpaint / generativeFill. Async — returns a task_id (auto-polled).playbackrate: >1 speeds up (2 = 2× faster), <1 slows down (0.5 = half speed). Async — returns a task_id (auto-polled).getTranscription with the returned id until job_status="DONE" (or "FAILED"). - Step 3: optionally call downloadTranscriptionFile to get VTT/SRT.target_task_id (UUID of the other task), relation_type (must be UPPERCASE): BLOCKS, RELATES_TO, or DUPLICATES. Use getTaskDependencies to read the current dependency graph before adding, to avoid duplicates. Returns the created relationship object.PROTECTED access.fields array. Common fields to copy: title, description, rich_description, assignees, tags, attachments, subtasks, priority, start_date, due_date, checklist. Optionally redirect the clone to a different status (column) via status. Returns {detail_url, task} — the cloned task object with expand=tags,attachments,subtasks.statuses field. primary_type is pm (default) or blogger — lowercase, no other values. access controls visibility: 0=PRIVATE, 1=PROTECTED, 2=PUBLIC.title. Optional: color (hex string e.g. #008CD3), order (integer position; appended at end if omitted). The new status id is returned and should be cached — it is required for createTask and searchTasks.status UUID is required and determines which board the task belongs to. Composite fields (assignees, tags) are handled by the toolkit's middleware — pass them inline and the correct sub-resource calls are made automatically after creation. Note: assignees are NOT reflected in the create response; use getTask?expand=assignees to verify.cloneBoard first if you want a backup. This action is not reversible via the API. Path Parameters: - board_id: Board UUID or slug.moveStatus to bulk-migrate tasks to another column before deleting.?expand=statuses to include the status columns inline; use ?expand=creator,modifier for user details. Path Parameters: - board_id: Board UUID or slug.?expand=creator for user details. Useful to confirm a status exists or to read its current order before repositioning.?expand= to pull nested data inline: assignees (user objects), tags (name list), subtasks, attachments, time_tracking, status_details, board. Multiple fields: ?expand=assignees,tags,time_tracking. Path Parameters: - task_id: Task UUID or slug.verb, actor, timestamp, and a diff of what changed. Useful for feeding AI context about recent task history.{blocks, blocked_by, related, is_blocked} — each list contains the counterpart task's {relationship_id, id, slug, title, complete, status}.listStatuses, searchTasks, and cloneBoard. Use this first when the user mentions a board by name — match on title from the response. Filter tips: search=<term> for server-side full-text match on board title/description (DB-backed — prefer this over fetching every board and filtering client-side); primary_type=pm (default) for task boards; is_public=true for public boards only; creator=<user_id> to scope to a creator; modified_after/modified_before for date-range filtering on last modification time (ISO 8601 date, e.g. 2026-01-01). Use page_size=50 to avoid missing boards on paginated workspaces. Always prefer this over searchBoards (removed) — ES index was unreliable.id values here are required by createTask, searchTasks, and moveStatus. Use ?expand=creator for user details.parent field points to this task. They follow the same schema as regular tasks. Use getTask?expand=subtasks for an inline alternative.updateTaskAssignees. Response: {"options": [{"value": <user_id>, "label": "Name (email)", "image": "..."}]} The value field is the integer user ID needed for updateTaskAssignees. Supports ?search=<name_or_email> for filtering.status_id (UUID of the destination status).relationship_id. Get the relationship_id from getTaskDependencies — it is the UUID on each entry in the blocks, blocked_by, or related arrays.-modified (newest first). Use this when you need "what changed recently" without knowing a specific board/status. For targeted search within a board column, use searchTasks instead — it supports richer filters (assignees, tags, priority, complete) and is what the UI uses for all board-view loading.board + status: The UI always passes board=<board_id> and status=<status_id> together, fetching one column at a time (page_size=10). Without scoping, results span the entire workspace (800+ tasks) with no relevance ranking. Correct agent pattern: 1. Call listStatuses to get status IDs for the board. 2. Call searchTasks?board=<id>&status=<sid>&search=<term> per column. Supports nested filters on assignees/tags, date ranges, and full-text search across title/description/tags/comments. SUBTASKS ARE EXCLUDED BY DEFAULT — pass include_subtasks=true to search them. A subtask that is absent from every result is not missing or unindexed; it is filtered out. Confirm with getTask by slug before concluding a ticket does not exist.title, description, access, extra (arbitrary JSON config merged into the board's _config). Path Parameters: - board_id: Board UUID or slug.order (integer) to reposition the column in the board view — other columns shift automatically. Updatable fields: title, description, color, order.status with the destination status UUID. Composite fields assignees_add/assignees_remove and tags_add/tags_remove are handled by the middleware — they trigger sub-resource calls automatically. description is converted to a Quill Delta (rich_description) by the toolkit. Path Parameters: - task_id: Task UUID or slug.listWorkspaceMembers to look up user IDs before calling this. Body: {"add": [<user_id>, ...], "remove": [<user_id>, ...]} Either key may be omitted. Returns {} on success (no body — by design). To verify or read current assignees: getTask?expand=assignees. Note: ES index updates async (~5s lag) — searchTasks?assignees=<id> may not reflect a new assignment immediately.getTask?expand=attachments). Body: {"add": ["<attachment_id>", ...], "remove": ["<attachment_id>", ...]}. Either key may be omitted. Returns the updated attachment list.createSocialMediaReviewBundle, without recreating it. The operation is idempotent — a draft already present in the bundle is silently skipped, not duplicated. Draft IDs come from getSocialMediaDrafts. Returns the updated bundle, including the full list of drafts and the shareable review URL.comments: a list of { "message": string, "delay": seconds } objects published after the post goes live, in array order. For a LinkedIn document (carousel) post, pass the PDF as additional.linkedin.document (see LinkedinDocument) with media: [].draft_ids to seed the bundle on creation, or create an empty bundle and append drafts later via addDraftsToSocialMediaReviewBundle. Draft IDs come from getSocialMediaDrafts. The response includes a linkToReview URL to share with reviewers.listSocialMediaTags first and reuse an existing tag if a tag with the same name already exists, to avoid creating duplicates.network when the user explicitly named exactly ONE platform (e.g. "show me my Instagram accounts" → network="instagram"). FORBIDDEN PATTERN — do not call this tool multiple times with different network values to "collect all platforms". That is always wrong: the unfiltered call already returned them.product=social_media so PM tags are not returned. Pagination is disabled.Click any prompt to copy it.
App Stats
116
Tools
3
Prompts
ChatGPT
Platforms
Category
Create VisualsWorks with
Similar Apps
Data refreshed daily
