MCP App Store
by KeenEthics
Unsora icon

Unsora MCP Server for ChatGPT

By Unsora

Overview

Unsora is an AI clip maker and social media scheduler. Paste a YouTube link, get captioned shorts back, and schedule them to TikTok, Instagram and YouTube without leaving the chat. Clip YouTube videos into shorts • Paste a YouTube link or a video file: a podcast, stream, webinar or interview • Get up to 20 clips, reframed to 9:16 for TikTok, Reels and YouTube Shorts, or 1:1, 4:5 and 16:9 • Add animated captions in 35 styles, and translate them • Let Unsora pick the strongest moments, or ask for specific ones, like "every product demo" Schedule social media posts • Schedule to TikTok, Instagram, YouTube, Facebook, LinkedIn and more from one post composer • Post to up to 10 accounts at once, with a different caption for each • Set YouTube titles, TikTok privacy and Pinterest boards, then post now, schedule or save a draft • Ask what's in your queue this week, and see views, likes, comments and shares for published posts Create the content too • Generate short AI videos for Reels, TikToks and ads • YouTube thumbnails from your photo and a style you like, talking-avatar videos, music and voiceovers • Upscale images to 8K and video to 4K Clips, videos and posts appear in interactive previews in the chat, and every file is saved to your Unsora library. You review the caption, media, accounts and time in the composer before anything is scheduled. You need an Unsora account to use this plugin. Clipping and generations spend credits from your Unsora account, and Unsora only accesses the social accounts you have connected yourself.

How to connect

  1. ChatGPT: Open the Unsora listing in the ChatGPT app directory and choose Connect. ChatGPT listing Unsora exposes 49 tools here.

Tools

audio_status

ChatGPT
Single-shot generation status check (used by the preview UI).

change_voice

ChatGPT
Re-voice existing speech into one of the user's cloned voices (Voice Changer, speech-to-speech): keeps the timing and delivery of the source, swaps the voice. The source can be audio or video, up to 5 minutes. Costs about 10 credits per minute. voiceId must be a cloned voice id from list_voices (create one with create_voice_clone). Poll with wait_for_voiceover. Renders a live audio player in app-capable hosts.

clipping_status

ChatGPT
Single-shot AI clipping job status + clips (used by the preview UI).

complete_upload

ChatGPT
Add a file uploaded through create_upload_url to the user's Unsora library. Returns the upload with its url, which works anywhere a media URL is accepted (create_post media, reference images, clipping input). In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

compose_post

ChatGPT
Open the interactive post composer: the user picks accounts, edits the caption and media, sets per-platform options (YouTube title/visibility/category/tags, TikTok privacy/comments/duet/stitch/branded content/AI label, Pinterest board/title/link, Google Business call-to-action button, video cover) and posts now, schedules or saves a draft — the composer creates the post itself. Use it whenever the user wants to create or schedule a post and the host can show apps; prefill everything the user already gave (caption, media URLs, accounts, time). After it opens, do NOT call create_post yourself; tell the user to finish in the composer. In hosts without app support, gather the details in chat and call create_post instead.

create_avatar_video

ChatGPT
Turn a portrait into a talking-head video (AI Avatar Maker, SkyReels V3): the person in imageUrl speaks a transcript (voiced with voiceId) or lip-syncs a supplied audioUrl. Costs 10 credits. Use list_voices to pick a voice — MiniMax presets, Eleven v3 voices, or the user's cloned voices. Poll with wait_for_video. Renders a live preview in app-capable hosts.

create_clipping

ChatGPT
Turn a long video into short clips (AI clipping). Use this for any request to clip, cut, chop, trim or repurpose a long video, podcast, stream, webinar or interview into shorts, Reels, TikToks or YouTube Shorts, or to find and extract specific moments or highlights (set query to describe the moments in natural language). Provide a public videoUrl. DO NOT call this tool until the user has explicitly chosen: aspect ratio, captions on/off (and style if on), number of clips, and clip length (not needed when query is set). This tool spends the user's credits — if any of these choices is missing from the conversation, you MUST stop and ask the user for all missing ones first. Never assume defaults, never infer unstated preferences. In app-capable hosts this renders a live preview that polls progress and displays the clips by itself — when the preview is rendered, do NOT call wait_for_clipping or clipping_status afterwards; summarize the job settings and end your turn. Only poll with wait_for_clipping in hosts without the preview.

create_image

ChatGPT
Generate an image for a social post, thumbnail or ad creative. Models include GPT Image 2.5 Flare / Sunburst, GPT Image 2, Nano Banana Pro / 2 / 2 Lite and Seedream 5.0 Pro — see list_models for each model's inputs. Attaching images in inputs edits / uses them as references. Priced live (get_price). Returns generation.id — poll with wait_for_image. Renders a live preview in app-capable hosts.

create_image_and_schedule

ChatGPT
Generate an image for a social post and schedule it in one step: create image → wait → schedule slideshow post to accounts. If any target account is Instagram, prefer aspectRatio 4:5 (or 1:1); other ratios are padded to fit. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

create_influencer

ChatGPT
Generate an AI influencer portrait for a social account or UGC-style post. Renders a live preview grid in app-capable hosts; poll each id with wait_for_image otherwise.

create_motion_control

ChatGPT
Make a character copy the motion of a reference video: the character in inputs.image performs the movement, dance or acting from inputs.video. Models: Kling 3.0 Pro / Standard, Kling 2.6 Pro, Wan 2.2 Animate, DreamActor v2 (see list_models). Priced live by the motion video's length — call get_price first. Poll with wait_for_video. Renders a live preview in app-capable hosts.

create_movie_material

ChatGPT
Generate pre-production reference images for AI films and ads (Movie Materials Generator, GPT Image 2): character face references, full-body references, turnaround sheets, location references, video first frames, style mood boards, and 2x4 / 1x4 storyboards. Feed the results to create_video as reference images for consistent characters and locations. Costs 3 / 4 / 6 credits at 1k / 2k / 4k. Poll with wait_for_image. Renders a live preview in app-capable hosts.

create_music

ChatGPT
Generate a music bed or song for a social video, Reel or ad (Mureka AI, song or instrumental BGM). Song models (auto, mureka-9, mureka-8, mureka-o2, mureka-7.6) require lyrics; mureka-7.5 generates instrumental BGM and treats lyrics as optional. The prompt sets genre, mood, tempo, and vocal style (e.g. 'r&b, slow, passionate, male vocal'). Returns generation.id — poll with wait_for_music. Renders a live audio player in app-capable hosts.

create_post

ChatGPT
Create, schedule, queue, publish or cross-post content to connected social accounts. Use this for any request to post now, schedule for later, publish to multiple platforms at once, or add something to the posting queue. Supported platforms: YouTube, TikTok, Instagram, Facebook, LinkedIn, Pinterest, Threads, Bluesky, X and Google Business Profile. When the user wants to set the post up themselves or per-platform options are still open (TikTok privacy, YouTube title, Pinterest board), prefer compose_post, which opens an editable composer in app-capable hosts. Without scheduled_at the post is saved as a DRAFT — set publishNow: true to post immediately. Requires paid plan. Images are fitted to each platform automatically at publish (format, size, Instagram's 4:5–1.91:1 ratio by padding, consistent carousels); 4:5 images look best on Instagram, 9:16 suits TikTok and Reels. Posts that break a rule Unsora can't fix (video length/size, image count, caption length) are rejected with an issues list per platform. YouTube only takes video and uses title (video title). Pinterest needs media (1 image, 2–5 images or a video). X takes text, 1 image, 2–4 images or 1 video (up to 140 s, 512 MB) with a caption of at most 280 characters. Google Business Profile takes text or exactly 1 JPG/PNG image (max 5 MB), no video, summary max 1500 characters, with an optional call-to-action button (google_business). In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

create_thumbnail

ChatGPT
Generate a YouTube thumbnail (16:9) for a video you are publishing. Returns one job per variation. Renders a live preview grid in app-capable hosts; poll each id with wait_for_image otherwise.

create_upload_url

ChatGPT
Get a one-time link for uploading a large or local file straight to the user's Unsora storage (no base64, far beyond upload_file's limits). Returns uploadUrl and blobName. PUT the raw file bytes to uploadUrl with the file's Content-Type, e.g. curl -X PUT -H "Content-Type: video/mp4" --data-binary @video.mp4 "<uploadUrl>", then call complete_upload with blobName to add it to the library and get its URL. Only useful when you can send HTTP requests yourself (e.g. a coding agent with a shell). In chat hosts that show apps, call list_uploads instead and ask the user to pick the file with the panel's Upload button.

create_video

ChatGPT
Generate a short video for a social post, Reel, TikTok or ad. Models include Seedance 2.5 / 2.5 Turbo / 2.0, Veo 3.1 (Fast, Lite), Gemini Omni 1.1 Flash, Kling O3 Pro / 3.0, Wan 3.0 and Grok Imagine Video 1.5 — see list_models for modes and inputs. Video is expensive: call get_price first, tell the user the credits, and pass them as maxCredits. Poll with wait_for_video. Renders a live preview in app-capable hosts.

create_voice_clone

ChatGPT
Create an instant voice clone from a speech recording (ElevenLabs). Only clone the user's own voice or a voice they have permission to use. Best results: 1–2 minutes of clean speech with no music or background noise. Costs 15 credits; max 10 clones per account. Returns clone.id — use it as voice_id in create_voiceover, create_avatar_video and change_voice. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

create_voiceover

ChatGPT
Generate a voiceover for a social video, Reel or ad from a script. Use list_voiceover_voices first and let the user pick an ElevenLabs Eleven v3 voice from the previews; the user's cloned voices and MiniMax presets (see list_voices) work too. Costs: Eleven v3 6 credits per started 1,000 characters; cloned voices 3 credits per started 500 characters; MiniMax presets 1 credit per started 1,000 characters. Supports <#x#> tags between words to pause for x seconds (0.01–99.99). Returns generation.id — poll with wait_for_voiceover. Renders a live audio player in app-capable hosts.

delete_generation

ChatGPT
Permanently delete one result from the user's library (type + id from list_generations), or an uploaded file (type upload, id from list_uploads). Confirm with the user first — this cannot be undone. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

delete_post

ChatGPT
Delete a post. Does not remove already-published content from the platforms. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

delete_voice_clone

ChatGPT
Permanently delete one of the user's cloned voices. Confirm with the user first. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

get_accounts

ChatGPT
List connected social accounts for scheduling. Supported platforms: YouTube, TikTok, Instagram, Facebook, LinkedIn, Pinterest, Threads, Bluesky, X and Google Business Profile (one account per business location). Each account has a provider field identifying its platform (YouTube accounts have provider "google"). In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

get_credits

ChatGPT
Get current credit balance. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

get_post

ChatGPT
Fetch one post by id, including per-account publish status and media. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

get_post_analytics

ChatGPT
Performance of the user's published posts (Scheduler Analytics): totals for views, likes, comments and shares plus per-platform breakdowns over the last N days. Pass refresh: true to pull fresh metrics from the platforms first (slower). In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

get_price

ChatGPT
Read-only price quote: returns the credit price of one video, image or motion-control generation for these exact settings (resolution, duration, audio, attached media length). It creates nothing, spends no credits and changes nothing in the user's account. The quoted credits can be passed as maxCredits to the create tool.

get_subscription

ChatGPT
Get plan and Stripe subscription state (check isActive before scheduling posts). In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

image_status

ChatGPT
Single-shot generation status check (used by the preview UI).

list_generations

ChatGPT
Browse the user's past results for one feature, newest first, with status and output URLs: image, video, music, voiceover, influencer, thumbnail, clipping, image_upscale, video_upscale, motion_control, avatar, movie_material, voice_change. Use it to find something made earlier (e.g. to post or reuse it). Uploaded files are in list_uploads. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

list_models

ChatGPT
Every video, image and motion-control model with its modes and exact input fields (WaveSpeed's own names, enums, ranges and defaults). Call this before create_video / create_image / create_motion_control to pick a model and build inputs.

list_posts

ChatGPT
List the user's posts, newest first: drafts, scheduled, publishing, published and failed. Filter with status (DRAFT, SCHEDULED, PUBLISHING, PUBLISHED, PARTIALLY_PUBLISHED or FAILED). In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

list_uploads

ChatGPT
List the user's uploaded media assets (from upload_file or the app). In hosts that show apps the panel has an Upload button, so this is also how the user uploads a file from their device (any size): call it and tell them to use the button. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

list_voiceover_voices

ChatGPT
List the 20 ElevenLabs Eleven v3 voices available for create_voiceover, each with a description, gender, accent, and a short preview clip URL. In app-capable hosts (e.g. Claude.ai) this tool renders an interactive voice picker with tap-to-play previews — the user already sees every voice, so do NOT repeat the voices as a table, list, or summary in your reply; just tell the user to pick from the picker above. Only in hosts without the UI, share the preview URLs so the user can hear a voice before choosing it.

list_voices

ChatGPT
List every voice the user can use: MiniMax presets (presets), ElevenLabs Eleven v3 voices with preview URLs (elevenV3Voices), and the user's own cloned voices (clones — use the clone id). Voices work as voice_id in create_voiceover and create_avatar_video; change_voice needs a cloned voice. Also returns the clone limit (maxClones) and cloneCreditCost. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

pinterest_boards

ChatGPT
Boards a connected Pinterest account can pin to (used by the composer UI).

publish_post

ChatGPT
Publish a DRAFT or SCHEDULED post to its accounts right now instead of waiting for its schedule. Requires paid plan. Publishing runs in the background: this returns status PUBLISHING straight away. Call get_post after a minute or two to see each account's result; if some accounts fail, use retry_post. In app-capable hosts the panel tracks publishing live, so don't poll get_post there. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

retry_post

ChatGPT
Retry a FAILED or PARTIALLY_PUBLISHED post. Re-attempts only the accounts that failed; already-published accounts are never re-posted. Runs in the background: call get_post after a minute or two for the outcome (in app-capable hosts the panel tracks it live). In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

tiktok_creator_info

ChatGPT
TikTok creator info for a connected account: allowed privacy levels, whether comments/duets/stitches are disabled, max video length (used by the composer UI).

update_post

ChatGPT
Edit a DRAFT or SCHEDULED post (published posts cannot be edited). Only the fields you pass change. Passing scheduled_at reschedules; passing scheduled_at: null converts it back to a draft. Passing accounts replaces the full target account list. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

upload_file

ChatGPT
Upload media to the user's Unsora library (available to every authenticated user). Pass a public source URL (preferred, up to 200MB) or a base64 payload for small files (≤ ~7MB). For larger local files use create_upload_url + complete_upload, or in hosts that show apps call list_uploads and have the user pick the file with the panel's Upload button. The file is stored and recorded as an upload asset; the returned url can be used anywhere a media URL is accepted (create_post media, reference images, clipping input). When the user attaches a file in the chat, pass it as file. In app-capable hosts the result renders as an interactive panel the user can see and act on — do NOT repeat its contents as a list or table in your reply; add only what the panel doesn't say.

upscale_image

ChatGPT
Upscale an image to 2K, 4K or 8K (Image Upscaler). Costs 2 / 3 / 5 credits. Pass a public imageUrl — a generation outputUrl or an upload_file url. Poll with wait_for_image. Renders a live preview in app-capable hosts.

upscale_video

ChatGPT
Upscale a video's resolution and sharpness (Video Upscaler). Models: standard (10 credits), ultra-1080p (6 / 11 / 16 credits for up to 5s / 10s / longer), ultra-4k (22 / 43 / 64 credits). Pass a direct video file URL (e.g. .mp4 — a generation outputUrl or an upload_file url). Poll with wait_for_video. Renders a live preview in app-capable hosts.

video_status

ChatGPT
Single-shot generation status check (used by the preview UI).

wait_for_clipping

ChatGPT
Poll an AI clipping job until COMPLETED or FAILED. Returns clips with output URLs. Do NOT use this when the live preview panel is rendered (it polls by itself) — only in hosts without the preview, or when the user explicitly asks for output URLs. Checks every 30 seconds until the job is COMPLETED or FAILED.

wait_for_image

ChatGPT
Poll an image job until COMPLETED or FAILED: create_image, create_influencer, create_thumbnail, upscale_image and create_movie_material. Checks every 30 seconds until the job is COMPLETED or FAILED.

wait_for_music

ChatGPT
Poll music status until COMPLETED or FAILED Checks every 30 seconds until the job is COMPLETED or FAILED.

wait_for_video

ChatGPT
Poll a video job until COMPLETED or FAILED: create_video, create_motion_control, create_avatar_video and upscale_video. Checks every 30 seconds until the job is COMPLETED or FAILED.

wait_for_voiceover

ChatGPT
Poll a voiceover or change_voice job until COMPLETED or FAILED. Checks every 30 seconds until the job is COMPLETED or FAILED.

Capabilities

WritesInteractive

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/unsora"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/unsora)