VibeFlow MCP tools - VibeFlow Canvas

These are the tools exposed by the VibeFlow MCP server. They’re available to any MCP-compatible client connected via Connect your agent. Read-only tools carry MCP readOnlyHint annotations, so clients like Claude run them without per-call confirmation; outward-facing tools (schedule_post, publish_post) always prompt.

Guides (call these first)

`get_content_guide`

Returns the content-operator workflow: the post lifecycle, the platform-first rule, writing in the user’s voice or a voice persona, the X, Instagram and TikTok rules, LinkedIn polls, gap tracking, review comments, the visual paths (photos, canvas graphics, and Video Lab videos), and the schedule/publish confirmation contract. No arguments. Call once per session before creating or editing posts.

`get_canvas_guide`

Returns the canvas workflow — how to brief the design agent, review frames, and deliver graphics. No arguments. Call once before creating graphics.

Workspace

`list_organizations`

Lists the workspaces your account belongs to. Every content tool takes one of these ids as organization_id. No arguments.

`get_workspace_context`

Loads the workspace’s content context in one call.

Arg	Required	Description
`organization_id`	Yes	Workspace id (from `list_organizations`).

Returns

{ business_context, memories, viral_formats: { linkedin, x }, platforms, capabilities: { instagram, tiktok, x_articles }, agent_canvas_id }

. platforms lists where the workspace can post: always linkedin and x, plus instagram and tiktok when they’re enabled for the workspace. capabilities.x_articles is true when the workspace can write X Articles with create_x_article. agent_canvas_id is the canvas where designed post graphics live; use it for create_canvas_frame and set_post_graphic.

`list_post_options`

Lists the workspace’s publishing options: platforms, connected social accounts, and saved voice personas. Accounts resolve automatically when omitted from create_post, so this is only needed when the user names a specific account (“post as the company page”) or wants to write like someone else.

Arg	Required	Description
`organization_id`	Yes	Workspace id.

Returns

{ platforms, accounts: [{ platform, integration_id, author_urn, label, kind }], voices: [{ voice_persona_id, name, platform, handle }] }

Voice

Every post should be written in the voice from get_voice: the learned voice of whoever is posting, or a saved voice persona when the user wants to write like someone else.

`get_voice`

Loads the voice to write in. Call it before drafting. By default it returns the voice of the person or page posting for this user. Pass accounts when posting as a specific account so the voice matches that identity, or voice_persona_id to write like someone else.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`platforms`	No	The platforms you’re about to write for: `linkedin`, `x`, `instagram`, `tiktok`. Defaults to the workspace’s enabled platforms.
`voice_persona_id`	No	A saved persona (from `list_post_options` or `learn_voice_persona`).
`accounts`	No	`[{ platform, integration_id, author_urn? }]` from `list_post_options`, when not posting as the default account.

Returns { voice_kind, voice: { guides, patterns, example_posts }, workspace_voice, how_to_use, platforms, x_char_limit }, plus writing_as (the posting identity) or persona (the voice persona). example_posts holds up to 5 real posts to mirror for rhythm, length, openings and vocabulary; agents should never copy their sentences or claims. When the identity has no learned voice yet, how_to_use says so and the workspace brand voice applies.

`learn_voice_persona`

Learns how another person or brand writes from their public posts and saves it to the workspace voice library, so drafts can be written like them. Check list_post_options first: a persona that’s already saved only needs get_voice. Learning runs in the background and takes about a minute.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`profile`	Yes	A LinkedIn profile or company URL, an X profile URL, or the person’s full name (resolved via LinkedIn search).
`platform`	No	`linkedin` or `x`. Only needed for a bare handle.

Returns { voice_persona_id, name, platform, status: "learning", next_steps }. Call get_voice with the voice_persona_id; until learning finishes it returns a “still learning” message.

A persona is inspiration for the drafts the user asked for, never the user’s own voice. Facts still come from the workspace, not from the persona’s posts.

Posts

`create_post`

Creates a post as a draft: one post with one variant per platform, each carrying real copy written for that platform. Never publishes or schedules. The platform choice belongs to the user, so agents shouldn’t add platforms nobody asked for.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`variants`	Yes	1–4 entries, one per platform. Fields below.
`gaps`	No	Up to 10 `{ kind: "photo" \| "data" \| "other", message }`: what the post still needs from the user. Saved with the draft and shown as an alert banner in the studio.
`title`	No	Short internal working title; omit for an automatic one.
`brief`	No	Internal note on the post’s angle.
`canvas_id`	No	Backing canvas when the post’s graphics live there.

Each variant:

Field	Required	Description
`platform`	Yes	`linkedin`, `x`, `instagram`, or `tiktok`. Instagram and TikTok only when `get_workspace_context` lists them in `platforms`.
`text`	Yes	The full post copy for that platform; for Instagram and TikTok, the caption. Finished copy, no placeholders.
`caption`	No	Caption, where the platform uses one.
`photo_suggestion`	No	The exact real photo that would fit (who or what, setting, framing). Saved on the draft and shown in the studio.
`integration_id`	No	Connected account to publish as (from `list_post_options`); omit for the default.
`author_urn`	No	LinkedIn author identity when the account has several.
`linkedin_poll`	No	LinkedIn only: publish the variant as a native poll. See LinkedIn polls.

Returns { post_id, post_url, variants: [{ variant_id, platform, photo_suggestion? }], status: "draft", missing_info? }. post_url opens the draft in the VibeFlow studio. X copy is weighted-counted server-side (URLs 23 chars, emoji 2) against the connected account’s limit: 280, or 25k with X Premium. A too-long tweet fails the whole call with a clear message; shorten and retry.

Platform rules

Platform	Copy	Visual
LinkedIn	Post text.	An authentic photo is the default. If none fits yet, set `photo_suggestion` and declare a `photo` gap.
X	Weighted 280 or 25k character limit, per account.	Optional.
Instagram	Caption up to 2,200 characters: hook in the first line, hashtags at the end, no links.	Required before scheduling: photos (2–10 publish as a carousel), one MP4 (a reel), or a 1080×1350 graphic.
TikTok	Same caption rules as Instagram.	Required before scheduling: one vertical 9:16 video (MP4, MOV or WebM), 1–35 JPEG/WebP photos as a photo post, or 1080×1920 graphics.

TikTok posts also need a privacy level and compliance settings (comments, duet, stitch, commercial disclosure, AI-generated label) that only the user can choose, in the studio’s TikTok tab. schedule_post and publish_post refuse a TikTok variant until they’re set.

LinkedIn polls

Pass linkedin_poll on a LinkedIn variant to publish it as a native poll.

Field	Required	Description
`question`	Yes	Up to 140 characters.
`options`	Yes	2–4 distinct answers, up to 30 characters each.
`duration`	No	`ONE_DAY`, `THREE_DAYS` (default), `SEVEN_DAYS`, or `FOURTEEN_DAYS`.

The variant’s text is the body above the poll; don’t repeat the options in it. A poll post is text only (no photos or graphics, so no photo gap), and the poll locks once the post is live. Polls exist only on LinkedIn.

`create_x_article`

Creates a long-form X Article as a draft. Never publishes or schedules. Only works on workspaces where get_workspace_context reports capabilities.x_articles: true; elsewhere the call fails. Use it when the user asks for an article, not as an extra next to a regular X post.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`article_title`	Yes	The headline shown on X, up to 200 characters. Don’t repeat it as a heading at the top of `body`.
`body`	Yes	The full article in markdown, up to 100,000 characters: `#` and `##` headings only (X renders two levels), bold, lists, links, and images.
`title`	No	Internal working title for the Posts list; defaults to `article_title`.
`brief`	No	Internal note on the angle.
`integration_id`	No	Connected X account to publish as (from `list_post_options`); omit for the default.
`photo_suggestion`	No	The real photo that would make a good cover, shown in the studio.
`gaps`	No	Up to 10 `{ kind, message }`, as on `create_post`.
`canvas_id`	No	Backing canvas when the cover graphic lives there.

Returns

{ post_id, post_url, variant_id, platform: "x_article", article_title, status: "draft", photo_suggestion?, missing_info? }

. The cover is the variant’s first visual: attach it with set_post_photos or set_post_graphic (X recommends 5:2; video isn’t supported). Inline images need hosted urls, so upload local files with request_upload_url first. Tweet character limits don’t apply. Edit the headline or body with update_post_variant (article_title, text), and schedule or publish with platforms: ["x_article"].

`list_posts`

Lists the workspace’s posts, newest first, with per-variant platform, lifecycle, schedule time, and whether a visual and connected account are set.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`limit`	No	1–50.
`status`	No	`scheduled` for the upcoming calendar only.

`get_post`

Reads one post in full: per-variant copy, visual_urls (photos), frame_ids (canvas graphics), media_layout, photo_suggestion, linkedin_poll, schedule state, missing_info gaps, and the post_url that opens it in the studio. Always call before editing, attaching visuals, scheduling, or publishing.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	From `list_posts` or `create_post`.

`update_post_variant`

Edits one variant of a draft or scheduled post: its copy, its photo_suggestion, its LinkedIn poll, or whether it’s enabled. Character limits are enforced server-side, and a too-long edit fails as a whole. Published variants only accept text edits. Visuals go through the two set_post_* tools instead.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.
`variant_id`	Yes	From `get_post`.
`text` / `caption`	No	New copy. On an `x_article` variant, `text` is the full markdown body.
`article_title`	No	`x_article` variants only: the headline, required before the article can be scheduled or published.
`enabled`	No	Turn this platform variant on or off.
`photo_suggestion`	No	The suggested real photo; an empty string clears it.
`linkedin_poll`	No	LinkedIn only: set or replace the poll. Send the whole poll every time.
`remove_linkedin_poll`	No	`true` removes the poll.

Send at least one change per call.

`classify_posts`

Labels posts with a content-strategy bucket (get-noticed, build-trust, selling, or customer), the labels shown on the workspace’s Drafts board. Omit assignments to auto-classify every unlabeled post in one batch; posts that already have a label are never touched.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`assignments`	No	Up to 50 `{ post_id, bucket }` to label specific posts. Omit to auto-classify.
`overwrite`	No	Let `assignments` replace an existing label. Only when the user asks.

Returns { mode: "auto", classified, assignments, has_more } when auto-classifying (call again while has_more is true), or { mode: "explicit", updated, skipped } for explicit labels.

`flag_post_gaps`

Updates what an existing post still needs from the user, shown as an alert banner on the draft in the studio. New posts declare their gaps on create_post instead. Kinds: photo, data, other. Replace semantics: gaps becomes the post’s full open list; an empty array clears it. Photos attached via set_post_photos clear photo gaps automatically.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.
`gaps`	Yes	Up to 10 of `{ kind, message }`.

Review comments

Posts carry the same review comment threads teammates leave in the VibeFlow studio. Your agent can read them, leave its own feedback, reply, and resolve threads once the change is made. Comments never change the copy: edits go through update_post_variant.

`get_post_comments`

Reads the open review threads on a post. Each thread has a thread_id, its messages oldest first, and an optional anchor pointing at a passage of one variant’s copy. No anchor means the comment is about the whole post.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.

Returns

{ post_id, post_url, open_threads: [{ thread_id, anchor?: { variant_id, platform, quoted_text, quote_still_in_copy }, messages: [{ author, from_vibeflow_staff, body, created_at }] }] }

. quote_still_in_copy: false means the copy was rewritten after the comment, so it may be stale.

`comment_on_post`

Leaves feedback on a post as a review comment. Comments are posted under the connected user’s name and the whole workspace sees them. Three shapes:

body alone comments on the whole post.
body + variant_id + quoted_text anchors the comment to that exact passage, highlighted in the studio. quoted_text must be a verbatim substring of the variant’s copy as get_post returns it.
body + thread_id replies to an existing thread, and reopens it if it was resolved.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.
`body`	Yes	The feedback, plain text, up to 4,000 characters.
`variant_id`	With `quoted_text`	The variant the quoted passage comes from.
`quoted_text`	No	The exact passage the comment points at.
`thread_id`	No	Reply to this thread instead of starting a new one. Not combinable with `quoted_text`.

Returns { post_id, thread_id, is_reply, anchored }.

`resolve_post_comments`

Marks review threads as resolved. Call it after the change a thread asked for is saved, and only for the threads actually handled. Questions, discussions, and threads that need the user’s input stay open.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.
`thread_ids`	Yes	1–50 thread ids from `get_post_comments`.

Returns { post_id, resolved, not_found }.

Visuals

Three separate paths: real photos and videos (list_workspace_photos or request_upload_url → set_post_photos), designed canvas graphics (create_canvas_frame → set_post_graphic), and videos made by the Video Lab agent (create_video → set_post_photos with the render url, see Video). The last one applied wins.

`list_workspace_photos`

Lists the workspace photo library (the Assets page in VibeFlow), newest first: each photo or video’s url, name, kind, and a caption describing what it shows. When a library photo fits a post, attach its url with set_post_photos.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`limit`	No	1–200, default 50.

Returns { photos: [{ url, name, kind, caption? }], total }. kind is image or video.

`request_upload_url`

Returns a short-lived signed URL to upload one photo or video into the workspace. HTTP PUT the raw bytes to upload_url with exactly the returned headers (Content-Type and Content-Length are part of the signature). After the PUT, public_url is the file’s permanent address.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`file_name`	Yes	Original file name with extension.
`content_type`	Yes	MIME type. Images: png, jpeg, webp, gif, avif (≤20MB). Video: mp4, webm, mov (≤60MB).
`size_bytes`	Yes	Exact file size in bytes.

Returns { upload_url, method: "PUT", headers, key, public_url, expires_in_seconds }. The URL expires in 10 minutes; one call per file.

`set_post_photos`

Attaches real photos or videos to a post as its published visuals. Each url can be:

a photo from list_workspace_photos,
any public image or video link (downloaded and stored in the workspace),
a public_url from request_upload_url, after the file was PUT.

Urls that can’t be used come back in skipped_urls with the reason. A photo pasted into the chat has no url: the user can drop it on the draft in the studio (post_url) or add it on the Assets page. Replaces designed canvas visuals if the post had them, and clears open photo gaps.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.
`image_urls`	Yes	1–20 urls from the sources above.
`variant_ids`	No	Limit to these variants; default all enabled.
`replace_existing`	No	Replace instead of add.

Returns { post_id, variants: [{ variant_id, platform, image_urls, replaced_canvas_visuals }], skipped_urls, remaining_gaps }.

`set_post_graphic`

Attaches designed canvas frames to a post. Exports the frames to PNG fresh on every call — re-call after a graphic is edited. canvas_id defaults to the workspace’s agent canvas.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.
`frame_ids`	Yes	1–10 frame ids.
`canvas_id`	No	Canvas holding the frames.
`variant_ids`	No	Limit to these variants.
`linkedin_layout`	No	`grid` (default, multi-image) or `carousel` (swipeable PDF, 2+ frames).

Video

The Video Lab agent makes and edits videos: cutting and captioning footage, shorts, titles, motion graphics, music, and voiceover. These tools only work when Video Lab is enabled for the workspace; otherwise they return an error.

`create_video`

Starts an asynchronous agent job from a natural-language brief. Without project_id it creates a new Video Lab project, which counts toward the workspace’s video allowance. Pass the project_id of an earlier run to iterate on that video (“make the captions bigger”) instead of starting over. Consumes credits.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`prompt`	Yes	What to make or change: the platform, target length, mood, and every word that must appear on screen, verbatim.
`project_id`	No	Continue an earlier project (from `create_video` or `list_video_projects`).
`name`	No	Name for a new project, up to 120 characters. Defaults to the start of the prompt.
`format`	No	New projects only: `landscape` 16:9 (default), `vertical` 9:16 for TikTok, Reels and Shorts, or `square` 1:1.
`video_urls`	No	Up to 10 footage links (YouTube, Loom, Vimeo, Drive, X, LinkedIn, direct files), or the `public_url` from `request_upload_url` for a local file.

Returns { job_id, status: "queued", project_id, new_project, next_steps } immediately. One run per project at a time: a call on a project that’s still working returns an error naming the running job_id.

`get_video_creation`

Polls a job started by create_video. Each call waits up to 20 seconds for progress. Keep polling until status is completed or failed; a run usually takes a few minutes.

Arg	Required	Description
`job_id`	Yes	From `create_video`.

Returns

{ job_id, status, project_id, video_lab_url, final_text, failed_reason, renders: [{ render_id, kind, status, url, duration_seconds, poster_url }], running_jobs }

. status is queued, running, completed, or failed, and kind is final, preview, or harness. Put a finished render on a post with set_post_photos.

`list_video_projects`

Lists the workspace’s Video Lab projects, newest first, with each one’s latest finished render. Use it to find the project_id to continue with create_video.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`limit`	No	1–50, default 20.

Returns

{ projects: [{ project_id, name, width, height, for_post_id?, video_lab_url, latest_render: { render_id, kind, url, duration_seconds, poster_url } | null, created_at, updated_at }] }

Publishing

`schedule_post`

Schedules a post for delivery at a specific time. Only after the user explicitly confirmed the exact post and time: the tool refuses without user_confirmed: true.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.
`scheduled_at_iso`	Yes	ISO 8601 with timezone offset, at least a minute in the future.
`timezone`	Yes	IANA timezone.
`user_confirmed`	Yes	`true` only on an explicit yes.
`platforms`	No	Schedule only these platforms: `linkedin`, `x`, `instagram`, `tiktok`, `x_article` (X Articles written in the studio). The post’s other drafts wait. Omit to schedule every enabled draft.

Variants that are already live are never re-sent. Returns { post_id, post_url, title, variant_ids, scheduled_at, timezone }.

`publish_post`

Publishes a post right now (delivery starts within seconds). Same confirmation contract as schedule_post.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.
`user_confirmed`	Yes	`true` only on an explicit yes.
`platforms`	No	Publish only these platforms now; the post’s other drafts wait. Omit to publish every enabled draft.

“Post the LinkedIn version now, X tomorrow morning”: publish_post with platforms: ["linkedin"], then schedule_post with platforms: ["x"].

`unschedule_post`

Cancels a scheduled delivery — the post returns to drafts. Safe and reversible.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`post_id`	Yes	The post.

Analytics

Read-only views over the workspace’s indexed social history (built by Index history in the app’s Analytics page) and its tracked competitors. Competitors are added and removed in the app.

`get_content_analytics`

Aggregated performance of the workspace’s indexed posts.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`dimension`	No	`overview` (default), `category`, `media`, `team`, or `cadence` (best posting days/times).
`platform`	No	`linkedin` (default), `x`, or `instagram`.
`range_months`	No	1–36, default 12.
`include_reposts`	No	Default false.
`timezone`	No	IANA timezone for cadence bucketing (default UTC).

`get_top_posts`

The workspace’s real best-performing indexed posts — copy, engagement, category, media type, image urls — optionally filtered by category or media_type (image, video, carousel, document, poll, text). Takes organization_id, optional platform (linkedin, x, or instagram), and limit (1–20).

`list_competitors`

Lists tracked competitor companies and creators with company ids, enrichment state, and indexing progress. Takes organization_id. Resolve ids here before the two tools below.

`get_competitor_analysis`

Head-to-head comparison vs tracked competitors: share of voice, engagement share, monthly trend, category mix, frequency, authors, follower counts. Takes organization_id, optional competitor_company_id (self-vs-one; omit for aggregate), range_months, include_reposts.

`get_competitor_evidence`

Bounded evidence for one competitor: strategic profile, top posts, topics, formats, engagement. Takes organization_id, competitor_company_id, optional limit (1–15).

Performance and audience

The app’s Performance > Posts table and Audience Intelligence page. Read-only, and for business accounts only.

`list_post_performance`

Every published post for the connected accounts with impressions, likes, comments, shares, engagement, and engagement rate, plus window totals. Answers “how did my posts do this month” or “which post got the most impressions”.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`platform`	No	`linkedin` (default), `x`, `instagram`, or `tiktok` (videos published from VibeFlow; `account` doesn’t apply).
`account`	No	Account name or handle. Omit for all; the response lists valid names in `accounts`.
`search`	No	Free-text match on the post copy.
`sort`	No	`recent` (default), `impressions`, `likes`, or `engagement`.
`since` / `until`	No	`YYYY-MM-DD` window. `since` wins over `range_days`.
`range_days`	No	Rolling window when `since` is omitted. Default 365.
`limit` / `offset`	No	Page size 1–100 (default 25) and offset.

Returns

{ platform, window, total, offset, has_more, stats, accounts, posts: [{ text, url, published_date, account, impressions, likes, comments, shares, engagement, engagement_rate }] }

`get_follower_growth`

Follower history per connected account: start and latest counts, net change, and a daily series. On LinkedIn it also returns the posts published in the window, so follower jumps can be lined up with specific posts.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`platform`	No	`linkedin` (default), `x`, or `instagram`.
`account`	No	Account name or handle. Omit for all.
`since` / `until`	No	`YYYY-MM-DD` window. `since` wins over `range_days`.
`range_days`	No	Rolling window when `since` is omitted. Default 90.

Returns { platform, window, accounts: [{ account, start, latest, change, change_pct, points }], posts_in_window? }.

`get_audience_overview`

Who engages with the workspace’s posts: total engagers, relevance split, engagement types, seniority, company sizes, follower bands, countries, industries, top companies and job titles, and repeat engagers. Also lists the tracked account names the other audience tools take as account.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`platform`	No	`linkedin` (default), `x`, or `instagram`.
`account`	No	Tracked account name. Omit for all.

`search_engagers`

Searches the people who liked, commented on, or reposted the workspace’s posts, with their enriched profile (company, title, industry, followers, location) and relevance verdicts. Pass post to list who engaged with one post, plus that post’s audience counts.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`platform` / `account`	No	As in `get_audience_overview`.
`query`	No	Free text over name, headline, company, about, industry, and location.
`company` / `job_title` / `country`	No	Field filters.
`min_followers`	No	Follower floor.
`engagement_type`	No	Exact type, e.g. `Comment`, `Like`, `Repost`.
`account_relevance` / `person_relevance`	No	`Relevant`, `Maybe`, or `Not relevant`, for the company and the person.
`post`	No	A post URL or id.
`sort`	No	`relevance`, `followers`, `recent`, `name`, or `comments` (commenters first, only with `post`).
`limit` / `offset`	No	Page size 1–100 (default 25) and offset.

Returns

{ platform, total, offset, has_more, post?, post_audience?, engagers: [{ engager_id, name, headline, company, job_title, followers, location, engagement_type, last_comment, account_relevance, person_relevance, relevance_why }] }

`get_engager`

One engager’s full profile (about, company description, HQ, funding, role tenure, relevance reasoning) and every post they reacted to or commented on, newest first, with their comment text.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`engager_id`	Yes	From `search_engagers` or `get_audience_overview`.

`ask_audience`

Answers a natural-language question about the engagers, engagements, and tracked posts (the Audience page’s Ask AI), for joins and counts the filters can’t express: “which VPs at fintech companies commented more than twice”. Unavailable while profiles are still being enriched. For simple filters, search_engagers and get_audience_overview are faster.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`question`	Yes	3–1,000 characters.
`platform` / `account`	No	As in `get_audience_overview`.
`post`	No	A post URL or id to scope the question to.

Returns { platform, title, count, rows, truncated?, sql }: up to 50 answer rows and the read-only SQL that produced them.

Radar

The app’s Radar page: what’s moving on the topics the workspace watches, and in its space. Radar has three feeds:

Feed	What it holds	Read with
Your topics	Signals the scheduled sweeps of the watched topics kept, each scored 0–100 for the business. Includes Trending on X when it’s on.	`list_radar_signals`
For you	Trends picked for the business on LinkedIn and X. Switched on per workspace by the VibeFlow team.	`list_radar_trends`
Timeline	LinkedIn and X posts worth engaging with, with drafted comments. Switched on per workspace by the VibeFlow team.	`list_radar_timeline`

The feeds fill on their own schedules, so there’s nothing to trigger. search_radar is the instant path: it scans any topic right now. Radar tools are for business accounts only.

`get_radar`

The starting point for any Radar question: the watched topics, the topic limit, whether Trending on X is on, and which feeds are switched on, with their counts and last and next run.

Arg	Required	Description
`organization_id`	Yes	Workspace id.

Returns

{ topics: [{ topic_id, topic, focus?, sources, subreddits?, cadence, enabled }], topic_limit, trending_on_x, for_you: { enabled, platforms, last_run_at, next_run_at, new_trends, posted_trends }, timeline: { enabled, platforms, last_run_at, next_run_at, new_posts, engaged_posts } }

. A feed that’s off comes back as { enabled: false }.

`list_radar_signals`

The Your topics feed, newest first. Each signal has a summary, why it’s notable, a 0–100 relevance score for the business, and the angle to take.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`topic_id`	No	Only signals from this watched topic (from `get_radar`).
`saved_only`	No	Only signals saved with the bookmark in the app.
`min_relevance`	No	0–100. 70+ is a strong match.
`since_days`	No	Only signals found in the last 1–90 days.
`sort`	No	`recent` (default) or `relevance`. Pair `relevance` with `since_days` for what’s fresh.
`limit`	No	1–100, default 20.

Returns

{ count, has_more, signals: [{ signal_id, topic_id, topic, source, title, url, summary, why_notable, relevance, angle, author, engagement, published_at, found_at, saved, idea_card_id? }] }

. idea_card_id is set when Radar already turned the signal into an idea card.

To turn a signal into a post, write it in the user’s voice (get_voice) from what the signal actually says, then call create_post.

`list_radar_trends`

The For you feed: trends picked for the business, each with a summary, why it fits, a post angle, a suggested format, momentum, a 0–100 fit score, and up to 3 real posts as evidence.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`status`	No	`new` (default), `posted` (already turned into posts), or `dismissed`.
`platform`	No	`linkedin` or `x`. Omit for both.
`limit`	No	1–60, default 10.

Returns

{ enabled, status, count, trends: [{ trend_id, platform, label, summary, why_for_you, angle, suggested_format, momentum, score, evidence: [{ url, author, text, engagement, posted_at }], found_at, status, used_post_id? }] }

. When For you is off for the workspace, it returns enabled: false with a message.

`list_radar_timeline`

The Timeline feed: recent LinkedIn and X posts worth engaging with. Each one comes with why to engage, a drafted comment and quote post (one per connected account when there are several), its metrics, a digest of its comments, and anything already done on it. Posts from the watch list come first. Read-only: commenting, liking, and reposting happen in the app.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`status`	No	`new` (default), `engaged`, `dismissed`, or `archived` (older or quieter posts that left New).
`platform`	No	`linkedin` or `x`. Omit for both.
`watched_only`	No	Only posts from people or topics on the watch list.
`limit`	No	1–50, default 10.

Returns

{ enabled, status, count, posts: [{ timeline_post_id, platform, url, author, text, published_at, metrics, score, why_engage, suggested_comment, suggested_quote, drafts_by_account?, can_repost, watch?, comments?, your_engagement?, found_at, status }] }

. When Timeline is off for the workspace, it returns enabled: false with a message.

`search_radar`

Scans a topic right now across web and news, LinkedIn, Hacker News, X, and Reddit, without waiting for a scheduled sweep. Results are raw: not saved to Radar and not scored for the business. Takes up to about a minute.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`topic`	Unless `topic_id`	What to scan, up to 200 characters.
`topic_id`	No	Rescan a watched topic with its saved focus, sources, and subreddits.
`focus`	No	An angle, e.g. “founder takes, not product launches”.
`sources`	No	Any of `exa` (web and news), `linkedin`, `hackernews`, `x`, `reddit`. Defaults to all five, or the watched topic’s sources.
`subreddits`	No	Up to 8 communities to limit Reddit to.

Returns { topic, focus?, sources, count, items: [{ source, title, url, summary, author, engagement, published_at }] }, up to 24 items.

`watch_radar_topic`

Adds a topic to the workspace’s Radar so it keeps getting swept, and its strongest finds land in list_radar_signals and the chat bot digests. A workspace watches at most 5 topics. A topic it already watches comes back with already_watching: true.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`topic`	Yes	2–120 characters: a theme, a competitor, a ticker.
`focus`	No	What counts, e.g. “only price-moving news”.
`sources`	No	Defaults to all five.
`subreddits`	No	Limit Reddit to these communities. Adds Reddit to the sources.
`cadence`	No	`daily` (default) or `anytime` (about every 2 hours).

Returns { topic_id, topic, already_watching, note }. The first sweep is queued right away.

`update_radar_topic`

Pauses or resumes a watched topic, or changes its focus, sources, subreddits, or cadence. Pass only what changes: an empty focus clears it, and subreddits: [] means all of Reddit. Paused topics keep their signals.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`topic_id`	Yes	From `get_radar`.
`enabled`	No	`false` pauses the topic, `true` resumes it.
`focus` / `sources` / `subreddits` / `cadence`	No	As in `watch_radar_topic`. `sources` and `subreddits` replace the current lists.

Returns { updated: true, topic }.

`remove_radar_topic`

Stops watching a topic and deletes the signals it collected. To keep the signals, pause the topic with update_radar_topic instead.

Arg	Required	Description
`organization_id`	Yes	Workspace id.
`topic_id`	Yes	From `get_radar`.

Returns { removed: true, topic }.

Memory

`remember`

Saves a durable user preference for the workspace’s content (voice, wording, topics). Comes back via get_workspace_context. Takes organization_id, text, optional kind (style / wording / fact), importance (1–5), and supersedes_id.

`forget`

Deletes a saved preference by memory_id. Takes organization_id and memory_id.

Canvas (graphics engine)

`list_canvases`

Lists your canvases, most recent first. Paginated.

Arg	Required	Description
`organization_id`	No	Scope to one workspace (from `list_organizations`).
`cursor`	No	Pagination cursor from a previous call.
`limit`	No	Max canvases per page. Default 50, max 200.

Returns { canvases: [{ canvas_id, name, organization_id, preview_image_url, created_at, updated_at }], next_cursor, is_done }.

`create_canvas`

Creates a new canvas.

Arg	Required	Description
`name`	No	Canvas name. Defaults to “Untitled canvas”.
`organization_id`	No	Workspace to create it in.
`design_system_ids`	No	Design systems to attach as canvas skills.
`selected_style`	No	A style preset key.

`get_canvas`

Returns one canvas’s metadata. Takes canvas_id.

`list_canvas_frames`

Lists every frame on a canvas (metadata, no HTML). Takes canvas_id, optional cursor and limit.

`get_canvas_frame_code`

Returns the raw HTML of one frame. Takes canvas_id and frame_id.

`screenshot_canvas_frames`

Renders 1–4 frames to images so the agent can see the actual result, not just its HTML. Takes canvas_id and frame_ids; returns the screenshots as inline images.

`create_canvas_frame`

Starts an asynchronous agent job: the VibeFlow design agent reads your natural-language brief, plans the frames, authors the HTML, and may create several frames in one run. Consumes credits. Also the edit path — brief it again on the same canvas, naming the frames to change.

Arg	Required	Description
`canvas_id`	Yes	The canvas id (`agent_canvas_id` for post graphics).
`prompt`	Yes	The brief: verbatim copy, purpose, platform, size, mood.
`design_system_id`	No	Load a specific design system for this run.
`organization_id`	No	Bill against this workspace’s credit pool.
`max_steps`	No	Max agent steps. Default 100, max 200.

Returns { job_id } immediately.

`get_canvas_frame_creation`

Polls an agent job started by create_canvas_frame. Takes job_id; returns the job status (queued / running / completed / failed) and, once completed, the created_frames (frame_id, name, width, height), the agent’s final_text, and its credit usage.

Canvas exports

`export_canvas`

Starts an export job for a canvas. Takes canvas_id, format (png, jpg, pdf, pptx, or all), optional frame_ids. Returns { job_id }.

`get_canvas_export`

Polls an export job. Takes job_id; returns the status and, once completed, one result per format with its download_url, filename, mime_type, and size_bytes.

Auth + errors

Every tool requires auth — OAuth (default) or an API key bearer header; see the overview. Errors come back as actionable messages, not bare status codes: an over-limit X post explains the weighted count, a photo url that can’t be attached comes back with the reason, a TikTok post without its studio settings says where to set them, and an empty balance returns an “insufficient credits” message instead of running. The performance, audience, and Radar tools need a VibeFlow business account, and the video tools need Video Lab enabled for the workspace; otherwise they return a message saying so.

Connect your agent

Claude, Claude Code, Cursor, Codex, OpenClaw, Hermes — step by step.

Credits

What each tool costs.

​Guides (call these first)

​get_content_guide

​get_canvas_guide

​Workspace

​list_organizations

​get_workspace_context

​list_post_options

​Voice

​get_voice

​learn_voice_persona

​Posts

​create_post

​Platform rules

​LinkedIn polls

​create_x_article

​list_posts

​get_post

​update_post_variant

​classify_posts

​flag_post_gaps

​Review comments

​get_post_comments

​comment_on_post

​resolve_post_comments

​Visuals

​list_workspace_photos

​request_upload_url

​set_post_photos

​set_post_graphic

​Video

​create_video

​get_video_creation

​list_video_projects

​Publishing

​schedule_post

​publish_post

​unschedule_post

​Analytics

​get_content_analytics

​get_top_posts

​list_competitors

​get_competitor_analysis

​get_competitor_evidence

​Performance and audience

​list_post_performance

​get_follower_growth

​get_audience_overview

​search_engagers

​get_engager

​ask_audience

​Radar

​get_radar

​list_radar_signals

​list_radar_trends

​list_radar_timeline

​search_radar

​watch_radar_topic

​update_radar_topic

​remove_radar_topic

​Memory

​remember

​forget

​Canvas (graphics engine)

​list_canvases

​create_canvas

​get_canvas

​list_canvas_frames

​get_canvas_frame_code

​screenshot_canvas_frames

​create_canvas_frame

​get_canvas_frame_creation

​Canvas exports

​export_canvas

​get_canvas_export

​Auth + errors

​Next

Connect your agent

Credits

Guides (call these first)

`get_content_guide`

`get_canvas_guide`

Workspace

`list_organizations`

`get_workspace_context`

`list_post_options`

Voice

`get_voice`

`learn_voice_persona`

Posts

`create_post`

Platform rules

LinkedIn polls

`create_x_article`

`list_posts`

`get_post`

`update_post_variant`

`classify_posts`

`flag_post_gaps`

Review comments

`get_post_comments`

`comment_on_post`

`resolve_post_comments`

Visuals

`list_workspace_photos`

`request_upload_url`

`set_post_photos`

`set_post_graphic`

Video

`create_video`

`get_video_creation`

`list_video_projects`

Publishing

`schedule_post`

`publish_post`

`unschedule_post`

Analytics

`get_content_analytics`

`get_top_posts`

`list_competitors`

`get_competitor_analysis`

`get_competitor_evidence`

Performance and audience

`list_post_performance`

`get_follower_growth`

`get_audience_overview`

`search_engagers`

`get_engager`

`ask_audience`

Radar

`get_radar`

`list_radar_signals`

`list_radar_trends`

`list_radar_timeline`

`search_radar`

`watch_radar_topic`

`update_radar_topic`

`remove_radar_topic`

Memory

`remember`

`forget`

Canvas (graphics engine)

`list_canvases`

`create_canvas`

`get_canvas`

`list_canvas_frames`

`get_canvas_frame_code`

`screenshot_canvas_frames`

`create_canvas_frame`

`get_canvas_frame_creation`

Canvas exports

`export_canvas`

`get_canvas_export`

Auth + errors

Next