> ## Documentation Index > Fetch the complete documentation index at: https://docs.gooseworks.ai/llms.txt > Use this file to discover all available pages before exploring further. # Goose Growth tools > Every tool the public GooseWorks MCP connector exposes to Claude, ChatGPT, and the CLI Connect Claude, ChatGPT, Claude Code, or any MCP client to the public GooseWorks MCP connector (`mcp.gooseworks.ai/mcp`) and it sees the tool catalog below. This is the connector's own catalog; the in-app coworker's tools are managed separately and may differ. Read-only tools never change anything; tools marked **billable** can spend credits when they generate media. ## Account & organization | Tool | What it does | Notes | | ---------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `account_update` | Save the legacy coworker preferences (company website, role, use cases, goals, freeform notes) when the user explicitly asks Goose to remember them. | | | `account_whoami` | One call for everything about the caller: user + token identity, the organization, every workspace scope the token can reach (the inputs for `file_*.scope`), the credit wallet and plan entitlements (`video_enabled`, `coworker_enabled`, \`… | read-only | ## Ad creatives | Tool | What it does | Notes | | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `add_campaign_concept` | Append ONE concept to a campaign (GOOSE-3345): a single small call that leaves every existing concept and its pending plans untouched. | | | `ads_approval_decide` | Act on a plan-mode creative: `approve` one creative (or every awaiting creative in a batch) — this is the billing gate; `revise` — recompose from a natural-language steer and/or a prompt-variant label (free; the plan goes back to composi… | billable | | `ads_approval_list` | Plan mode: the creatives whose plans are composing or awaiting review, with each frozen `plan` (composed prompt, reference images, quality, ratios) and `counts`. | read-only | | `ads_creative_edit` | Paid edits on one creative, one `action` per call: `regenerate` (re-roll a variation, or `mode: edit\|exact` against a specific render), `resize` (faithful reframe of an approved render into platform placements), `layerize` (split a rende… | billable | | `ads_creative_read` | Read generated ad creatives for one brand. | read-only | | `ads_creative_update` | Set a creative's organization tags (full set; `[]` clears) and/or record the user's feedback on one of its renders (happy/neutral/sad + comment + reason chips; `null` clears a field). | | | `ads_generate` | Generate branded static ads through the same cloud workflow the app uses (compose → generate → persist → judge). | billable | | `ads_library_scrape` | Import ads from the public Meta Ad Library: the brand's own ads (omit `source_id`; designed graphics become remixable templates and the copy is summarized into the kit) or one tracked competitor/own source (`source_id` from \`brand\_get\_co… | | | `ads_template_create` | Bring your own ad: register one of the org's own images as a private static template — remixable like the catalog but visible only to your org (or to the Community with `is_public`). | | | `ads_template_read` | Read source templates without changing them. | read-only | | `ads_template_update` | Patch one of the org's own templates (uploads, scraped, Community snapshots): replace its tag set (send the full list; `[]` clears), retitle, file it under a product, toggle Community sharing, or delete it. | | | `campaign_read` | Read GooseWorks campaign planning records without changing them. | read-only | | `campaign_upsert` | Create a campaign when campaign\_id is omitted, or update an existing campaign when campaign\_id is provided. | | | `propose_campaign_concepts` | Propose concepts for a campaign from the brand's own evidence (Meta performance, creative feedback, learnings, operating plan) and write them onto the campaign. | | | `request_campaign_generation` | Request creative generation for a campaign's concepts (each concept's templates × variants). | | | `update_campaign_concept` | Edit ONE concept by id — only the fields in the patch change (GOOSE-3345): wording, a template swap, or generation settings (variant\_count, ratios, engine, quality). | | ## Automations | Tool | What it does | Notes | | ------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `automation_delete` | Permanently delete an automation and its files. | | | `automation_list` | List the brand coworker's scheduled automations (cron, timezone, status, run counts, credits) plus its creative schedules. | read-only | | `automation_upsert` | Create an automation (omit `automation_id`; needs `name` + `cron_expression`, and `agent_prompt` for agent mode or `script_content` for script mode) or patch one — only the fields you pass change; `cron_expression`/`timezone` changes rec… | | ## Brands & brand kits | Tool | What it does | Notes | | ------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `brand_create` | Create a brand in the caller's organization. | | | `brand_get_context` | The single context-loading call for a brand: pick `sections` and get the summary row, the canonical brand kit (positioning, palette, typography, reference images, competitors, existing ads), the product catalog, tracked ad-library source… | read-only | | `brand_list` | List the organization's brands (lean rows, no brand kit), most recently updated first. | read-only | | `brand_onboarding` | Start, read, and advance the same resumable Growth onboarding used by the GooseWorks web app, directly inside ChatGPT, Claude, Cowork, or a coding agent. | | | `brand_update` | One typed patch for everything on a brand: the brand row, the canonical kit fields (positioning, voice, palette — each edit is recorded as a user override that research re-runs preserve), a research kit-patch + `finalize_research`, produ… | | | `competitor_link` | Start tracking one competitor against a brand from whatever the user gave you: a name ('Allbirds'), a website, an Instagram or TikTok profile URL, or a Meta Ad Library URL. | | | `competitor_read` | Read tracked competitor brands without changing research. | read-only | | `competitor_research` | Re-run automatic research on a competitor you already track: refresh their profile, follower counts and recent posts, re-check their ad library, and re-synthesize the structured profile. | | | `competitor_suggestions` | Competitors we already discovered for this brand during research but that the user has NOT linked yet. | read-only | | `competitor_unlink` | Stop tracking a competitor and drop the research gathered about them — their stored posts, follower history and scraped ads go with it. | | ## Chat & delegation | Tool | What it does | Notes | | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `chat_get` | One session's transcript, oldest first, paginated with `after` (ISO created\_at cursor); `include_tool_calls` adds raw tool-call JSON on assistant messages. | read-only | | `chat_search` | With `query`: lexical search across the brand coworker's past messages (ranked snippets with message/session ids, timestamps, role, source) inside a time window. | read-only | ## Data providers | Tool | What it does | Notes | | -------------------- | ----------------------------------------------------------------------------------------------------------------- | -------- | | `data_call_provider` | GET an official ScrapeCreators endpoint through the GooseWorks proxy — no provider key, no local CLI. | billable | | `data_post_provider` | POST to an official ScrapeCreators endpoint that requires a body (bulk search etc.) through the GooseWorks proxy. | billable | ## Delegated tasks | Tool | What it does | Notes | | ---------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `goose_get_task` | Poll a delegated task. | read-only | | `goose_run_task` | Delegate a multi-step job to the brand's Goose coworker — the same server-side agent that powers the in-app chat, with our prompts, runtime widgets and approval gates. | billable | ## Files | Tool | What it does | Notes | | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `file_delete` | Delete a file or a directory (recursive) from the workspace. | | | `file_list` | List one directory (not recursive; dirs first, then files with size and modified time) or, with `query`, find files anywhere in the workspace by path substring (case-insensitive, up to 200 matches, `path` scopes the search). | read-only | | `file_read` | Read one workspace file inline — text, or base64 for binary — up to 10MB. | read-only | | `file_url` | A short-lived presigned S3 URL: `mode: upload` → PUT (`curl -X PUT -T "$url"`, plus the `Content-Type` header when you set `content_type`; max 50MB advisory; skill paths refused); `mode: download` → GET (`curl "$url" -o `). | | | `file_write` | Create or overwrite a workspace file from an inline string (base64 with `is_base64` for binary), up to 10MB; intermediate directories are created. | | ## Job status | Tool | What it does | Notes | | ------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `job_cancel` | Cancel a queued job before it starts — today: a queued social post batch (held credits are released) and plan-mode ad batches that have not been approved. | | | `job_get` | Poll any async job by id: ad generation batches, renders, animate-image jobs, layer splits, product imports, photo generations, brand research, ad-library scrapes, social post batches/creatives and delegated Goose tasks. | read-only | ## Media library | Tool | What it does | Notes | | -------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `media_list` | List a brand's registered media, newest first, filtered by scope, kind, status or tags. | read-only | | `media_update` | Patch one media asset: title, tags, status (draft → review; approval is done by a human in the app), metadata, which scopes it is linked to, or delete it. | | | `media_upload` | Register a media asset under a brand, product, video project, post or social account, from one of three sources: a file you will PUT to the returned presigned URL, an already-public URL (stored verbatim — the server never fetches arbitra… | | ## Organic social | Tool | What it does | Notes | | ---------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `competitor_search_mentions` | SPENDS CREDITS. | billable | | `organic_creative_queue` | Queue an APPROVED organic creative to its account's linked destinations (1-2 owned TikTok/Instagram accounts) as a MANUAL to-do entry — nothing auto-publishes (GOOSE-3235). | | | `social_account_list` | List a brand's social accounts (brand and creator accounts) with their identity summary, platform, lifecycle stage and setup status (research / strategy / initial batch). | read-only | | `social_account_upsert` | Create a social account (omit `account_id`; needs `platform` + `display_name`, and `account_type` brand\|creator) or patch one. | | | `social_analytics_get` | Read performance for a brand's managed social accounts: an overview, per-account metric snapshots, and per-post metrics joined to the posts Goose produced (via placements). | read-only | | `social_inspiration_library` | Read the brand's saved Inspiration library: each saved post itself (url, permalink, caption/note, thumbnail, creator, content type), how it was saved (when, by whom, research vs competitor vs manual), and — where GooseWorks also tracks t… | read-only | | `social_inspiration_save` | Save an external post (URL + optional stats and note) to the brand's inspiration store (Intelligence → Inspiration, D30), optionally linked to one account so it seeds research and strategy. | | | `social_inspiration_search` | Search the brand's saved inspiration and the research runs' discovered posts by text, platform or account. | read-only | | `social_inspiration_watch` | Download one saved inspiration VIDEO and analyse it with ffmpeg, returning a structured summary — duration, container/codec/resolution, scene boundaries, extracted-frame timestamps and, where audio and a transcription key allow, a transc… | | | `social_post_create` | Generate `count` posts for an account in a given format through the server pipeline (evidence → copy → judge → pair assets → render), or create one hand-authored post. | billable | | `social_post_list` | List an account's posts newest first — the review queue (`status: review`), approved, discarded, drafts — each with its idea reference, linked creative (render status, output/thumbnail URLs, credits) and placements. | read-only | | `social_post_update` | Patch a post's content (caption, hook, hashtags, plug level, idea, media, rendered output, status, posted\_at) and/or run one review action: `approve` (hands it to managed publishing, D45), `discard`, `restore`, `retry` (re-render, credit… | billable | | `social_strategy_get` | The single context-loading call for one social account: identity, the active strategy version (concept, audiences, pillars, voice, rules, tactics, avoid), linked products, the full pillars → buckets (lifecycle, evidence, ideas) tree, and… | read-only | | `social_strategy_upsert` | Write the strategy tree: activate a new strategy version (pass `expected_active_version` to replace a changed one; identical retries are idempotent), and create/update pillars (by `id`, else `title`), buckets (by `id`, else `pillar_id`+\`… | | ## Product photos | Tool | What it does | Notes | | ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `photos_generate` | Shoot branded product photos through the studio pipeline (compose, generate, judge, retry; retries are free). | billable | | `photos_read` | Read product-photo work with one tool: generation\_id returns one generation and its outputs; omit it to list a brand's outputs with product, status, archive, and cursor filters. | read-only | | `photos_update` | `approve` links an output to the product and makes it available in the brand kit for ad generation; `archive` (with an optional reason, recorded as a brand "dont" learning) removes it from downstream generation. | | ## Templates & catalog | Tool | What it does | Notes | | ---------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `catalog_fetch` | Fetch one catalog item by slug: a skill with its full instructions, scripts, files and dependency skills (follow `content` and use the live tools it names; in a terminal-free client use the skill's MCP-native path); a format's registry r… | read-only | | `catalog_search` | Discover what Goose can do: `skill` — the GooseWorks skill catalog (the same one `gooseworks search` reads; slug, name, description, category); `format` — the format registry (server-rendered organic/ad formats with status and platform);… | read-only | ## Video projects & rendering | Tool | What it does | Notes | | ---------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------- | | `set_final_render` | Choose which version the app shows as the final ad for a project. | | | `submit_render` | Kick off a render for a video ad project. | | | `update_render_status` | Sandbox-side callback that moves a render between statuses and reports visible video QA progress. | billable | | `video_catalog_list` | Browse what a video project can be built from: `characters` and `worlds` (Higgsfield stock shared across orgs plus the org's own uploads), `voices` (the curated ElevenLabs allowlist, with search and gender/accent filters), and `formats`… | read-only | | `video_project_read` | Read video work with one tool: project\_id returns one project with optional assets, scenes, messages, or renders; batch\_id returns one concept batch; omit both to list a brand's projects. | read-only | | `video_project_upsert` | Create a video project or concept batch when project\_id and batch\_id are omitted; otherwise update an existing project and/or batch with patch. | | | `video_render_run` | Render a video project on the server (the Playwright/ffmpeg render worker, D15). | billable |