QynTro AI

By UDDHAB PRAMANIK

Overview

Create media from ChatGPT with the credits in your QynTro account. • Images from FLUX, GPT Image, Nano Banana and Stability AI models • Image editing with an instruction, including placing a logo or product from a second image • Video from text, or from a picture you just made; video editing • UGC spokesperson ads from one photo, lip-sync, and music videos cut to a song • Original music tracks up to two minutes • Voiceovers in many voices and languages • Talking avatars: a studio presenter or your own photo speaking a script • PowerPoint decks (.pptx) from a topic • Textured 3D models (GLB) from a photo or a description Everything you make is saved to your QynTro library. Sign in with your QynTro account; each generation uses credits from your existing balance, and every reply says what it cost. Tools that bring a real face to life ask you to confirm you have the right to use that face and voice.

Tools

check_credits

ChatGPT

Show the connected QynTro account's media credit balance.

edit_image

ChatGPT

Edit an existing image with an instruction. Four models: nano-banana-2-edit (Nano Banana 2.1 by Google, the default), gpt-image-2.5-flare-edit and gpt-image-2.5-sunburst-edit (OpenAI - suited to text inside the picture), seedream-v5-pro-edit (Seedream 5.0 Pro by ByteDance). Pass the image URL - typically the URL a generate_image or edit_image call just returned (this is how "make me a poster, now make the sky purple" works), or any public https image. SYNCHRONOUS: the edited image comes back in this reply, usually within 10-40 seconds. Spends media credits from the account. COST RULE: if the user did not ask for a resolution, OMIT it - the server defaults to 1K, the cheapest full size.

edit_video

ChatGPT

Edit an existing video with an instruction - change the look, the setting or the action of a clip. Two models: gemini-omni-edit (Gemini Omni by Google, the default) and flux-3-edit-video (FLUX.3 Edit Video - MP4 only, clips up to 15 seconds). Pass the video URL - typically the URL a finished generate_video, image_to_video or get_job result returned, or any public https video. Async like generate_video: returns a job id prefixed "e" (e.g. e12) - collect it with get_job. Flat price per edit from the account's media credits; the reply states it.

generate_3d

ChatGPT

Build a textured 3D model (GLB mesh) from a photo, from up to four views of an object, or from a description. Models: hi3d-image (one photo, 210 credits), hi3d-multiview (up to four views for a cleaner back, 210 credits), meshy-text (a description, 120 credits). Async: returns a job id prefixed "d" (e.g. d12) - collect it with get_job. Flat price per model from the account's media credits; the reply states it.

generate_avatar_video

ChatGPT

Make a video of an AI avatar speaking (HeyGen). Two ways: a STUDIO avatar (avatar_id from list_avatars) speaking a script; or a PHOTO (image_url) brought to life - the face in the picture speaks the script, or lip-syncs a recording given as audio_url (a generate_voiceover result, for example). Async: returns a job id prefixed "a" (a12) - collect it with get_job. Charged per second of speech from media credits: studio avatars 3 per second, photos 5 per second. With a script the voice is optional (a default is used - pass gender for a photo), or pass a voice_id from list_avatar_voices. Script max 1500 characters. With a photo, consent must be true (ask the user to confirm the right to use this face and voice), and a recording is measured from the file (up to 60 s).

generate_image

ChatGPT

Generate one or more AI images from a text prompt using QynTro AI. Spends media credits from the connected QynTro account. Leave model out unless the user names one - the default costs 2 credits an image. Returns the stored image URL(s).

generate_music

ChatGPT

Generate an original music track from a text description (ElevenLabs Music). SYNCHRONOUS: the call blocks 30-90 seconds and returns the finished audio URL - tell the user it is composing rather than retrying. Costs media credits by length bracket (a 30s track costs the least; the price rises at 30s/120s boundaries). Max 120 seconds. Nothing is charged on failure.

generate_music_video

ChatGPT

Cut a music video to a song (AI Music Video, PixVerse VibeMV). The song comes from the QynTro music library (track_id) or an https MP3/WAV/AAC URL of 10 seconds or more (a generate_music result URL works). 30 or 60 seconds at 720p or 1080p: 180 / 270 / 360 / 540 media credits; a longer track is cut to the chosen length. Fifteen looks (Cinematic default). Optional image_url = character reference: the face, hair and age carry across every shot; clothes and setting do not, they follow the look. Async: returns a job id prefixed "m" (e.g. m3) - collect it with get_job; it takes seven to nine minutes.

generate_presentation

ChatGPT

Create a real .pptx presentation from a topic and return a download URL. SYNCHRONOUS (about a minute for text; longer with per-slide images) - tell the user it is building. Costs 2 media credits per slide, +8 per slide if images are on. Default 6 slides, no images. Also opens (and stays editable) in the QynTro dashboard.

generate_ugc_video

ChatGPT

Make a 10-second spokesperson ad from ONE photo (UGC Video, Boreal): the person in the photo says the words, filmed where the photo was taken - THE PHOTO IS THE SET, so give it a photo composed as the scene you want. Speech, lip movement and room tone come back in the render; nothing is recorded. Async: returns a job id prefixed "i" (e.g. i12) - collect it with get_job (two to three minutes). Costs media credits (30 today, 1080p pinned). For a particular voice: make the line with generate_voiceover first and pass its URL as audio_url.

generate_video

ChatGPT

Start an AI video render from a text prompt using QynTro AI. Rendering takes 1-4 minutes, so this returns a JOB ID immediately - call get_job with it to collect the finished video. Spends media credits (video is charged per second, so it is far dearer than an image). IMPORTANT COST RULE: if the user did not name a model or a duration, OMIT those arguments - the server then picks the cheap model and a short 5-second clip on purpose. Only pass model/duration/resolution/audio when the user actually asked for them. Always tell the user how many credits it cost (the reply states it).

generate_voiceover

ChatGPT

Turn text into natural speech (ElevenLabs or OpenAI voices) and return the MP3 URL. SYNCHRONOUS - returns in a few seconds. Spends CHARACTER credits (1 per character of text), a SEPARATE balance from media credits. Name a voice or omit it for the default; call list_voices for the available names. ElevenLabs voices speak any language - just write the text in that language.

get_job

ChatGPT

Check an async render started with generate_video, image_to_video, edit_video, generate_avatar_video, generate_3d, generate_ugc_video, lip_sync or generate_music_video. Returns the finished file's URL when ready (a video, or a GLB model from generate_3d), or that it is still rendering. Pass the job id you were given; omit it to see the most recent jobs. If a render failed, the credits were refunded automatically.

get_upload

ChatGPT

Collect a photo the user uploaded through a get_upload_link link: returns its permanent https URL once the upload is done - pass that as image_url to generate_avatar_video, lip_sync, generate_ugc_video, image_to_video, edit_image or any tool that takes a picture - or says it is still waiting. Call it after the user says the photo is uploaded; do not poll. Omit upload_id to list the latest uploads.

get_upload_link

ChatGPT

Use this when the user wants to use their OWN photo - a selfie or a face for an avatar, a product photo, any picture they have - with a QynTro tool that takes image_url (generate_avatar_video, lip_sync, generate_ugc_video, image_to_video, edit_image and the others). A picture attached to this chat cannot be passed to those tools. So: 1) call get_upload_link; 2) give the user the link - they open it on their phone or computer, choose the photo (JPG, PNG or WEBP, up to 12 MB) and tap Upload; if you can read the file yourself, you may instead POST it to the same link as multipart field "file" (curl -F file=@photo.jpg <link>); 3) when the user says it is uploaded, call get_upload with the upload id for the photo URL. The link works once, for 30 minutes, and costs no credits. Not needed for pictures QynTro made - pass their URLs directly.

image_to_video

ChatGPT

Animate an existing image into a short AI video clip. Pass the image URL - typically the URL a generate_image call just returned (this is how "create an image, then animate it" works), or any public https image (PNG/JPG/WEBP). Async like generate_video: returns a job id prefixed "i" (e.g. i12) - collect it with get_job. Charged per second of output from the account's media credits. COST RULE: if the user did not name a model or duration, OMIT them - the server defaults to the cheap model and a 5-second clip.

lip_sync

ChatGPT

Make the face in a photo sing or speak an audio file (Lip-Sync Video, MiniMax H3 Max). Audio must be 5 to 30 seconds; longer is refused and nothing is charged. The length is MEASURED from the file and billed per second from media credits: 480p / 768p (default) / 1080p at roughly 5 / 8 / 16 credits per second, x1.2 when over 15 seconds (as the model bills), 30 minimum. Async: returns a job id prefixed "i" - collect it with get_job (two to four minutes). consent must be true: ask the user to confirm they have the right to use this face and this voice.