Adly

By Adly

Overview

Adly is an AI marketing platform for small businesses and agencies. Tell ChatGPT what you want to promote and Adly designs the post, carousel, video or ad on your brand (your logo, colors, fonts and tone), writes the copy in Arabic or English, and publishes or schedules it to your Facebook Page, Instagram, LinkedIn profile and X account. What you can do: - Create single-post designs, tip carousels, product-photo ads, UGC-style and scene-based videos. - Write and edit ad copy, then approve it in chat. - Check what is ready, preview it, and ask for changes. - Publish now or schedule for later, list upcoming posts and cancel one. - Open the Adly content board from the sidebar to see your latest content and everything that is scheduled. Limits to know: creating content counts toward your Adly usage limit. You need an Adly account to connect. Social accounts are connected once on getadly.com. Posts to X have no links, and X posting is not included in every Adly account. TikTok publishing is coming soon.

Tools

add_brand_screen

ChatGPT

Add a brand color/logo end card clip to the timeline. Same as UI brand screen.

add_uploaded_video_scene

ChatGPT

Add a user-uploaded video as a ready Scene Studio clip. Same as UI upload scene.

add_video_export_music_clip

ChatGPT

Add a selected music track onto the export music timeline.

add_video_layer

ChatGPT

Add a text or image overlay layer on the Scene Studio timeline.

add_video_scene

ChatGPT

Add a new pending AI scene to Scene Studio (max 6). Same as UI Add scene.

ai_write_video_prompt

ChatGPT

Same as UI “AI writes”: expand a short idea into a production-ready video brief AND decide voice narration on/off, AI Writes personality / narration_mood, narrator sex/age, and on-screen text options. Use surface=scene_studio for Scene Studio master briefs (then pass prompt + options into create_video_scenes), or surface=create_video for short create-new videos (then create_video). Mode use_idea keeps the user’s idea; invent ignores idea and invents from brand + personality. Prefer personality=auto unless the user picked one. For emotional VO use personality=emotional_delivery (aliases emotional/heartfelt also map here).

analyze_brand_style

ChatGPT

Analyze uploaded Teach Adly your brand style examples with AI and generate brand instructions. Requires at least 3 uploaded examples. Uses 1 text from the usage limit, same as the Adly UI Brand Power flow.

apply_brand_style_learning

ChatGPT

Apply analyzed Teach Adly your brand style instructions to the brand prompt_instructions field so future Adly generations use the learned visual patterns.

approve_copy

ChatGPT

Use this when the user is happy with the caption Adly wrote. This marks the post ready; it does not publish anything. Then ask where and when to post, and use publish_now or schedule_post.

approve_design

ChatGPT

Use this when the user likes a design and says to go ahead (for example "looks good" or "approve it"). Approves the chosen design option and continues processing: Adly then writes the ad copy. For multi_mode: first approve applies the template to all products (starts variant generation); after products are ready, approve again to generate ad copy. For asset_design jobs, completes the job with no ad copy. Two-k draft jobs upscale to 4K PNG on approval; four-k jobs approve in place. Never call approve_copy afterward for asset_design.

approve_music

ChatGPT

Choose the music track for one clip of an Add Music job. Adly mixes it into the clip; when every clip has a track the job is complete and get_job media holds the final video. Use that video_url with create_ad_copy to get the publish caption.

cancel_job

ChatGPT

Cancel stuck MCP-created jobs only (source_channel=mcp or creative_brief.source=mcp) to free creation slots. Non-MCP Adly app jobs are refused and left unchanged. Pass job_id or job_ids. Publish-ready MCP jobs (await_publish_approval) are marked completed; others rejected. Does not delete media; use get_job first if you need URLs.

cancel_scheduled_post

ChatGPT

Use this when the user wants to cancel an upcoming scheduled post before it goes out. Does not delete posts that are already live.

clone_talking_character

ChatGPT

Clone a saved cartoon/talking character into the same brand. Optionally rename, restyle, or regenerate its scene image.

clone_video_export_music_clip

ChatGPT

Duplicate an export music clip.

create_ad_copy

ChatGPT

Use this when the user wants to post a video or image exactly as it is (their own, one made in Adly, or one you made in this chat), or wants a caption or ad copy written for it, the same as the Ad Copy job in Adly. Upload it first (upload_product_image or upload_video) and pass the returned URL as media_url. Designs made with the create tools get their caption after approve_design, so do not use this for them. The job ends at await_copy_approval; approve_copy makes it ready to publish in Adly. Pass source_job_id when the media is the output of an Adly job (for example an Add Music job) so Adly reuses that job instead of transcribing. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_asset_design

ChatGPT

Use this when the user wants a logo, a print or physical item, or a classic graphic design piece: posters, flyers, A4/A5 papers, menus, shirts, banners, signage, car stickers, logos, packaging labels, business cards, shelf talkers, and similar (for example "design a logo for my company" or "make a menu for my cafe"). Not for social media posts: use create_general_design. Put item type, material/use case, purpose, and required visible copy inside the brief. YOU choose the best output size/ratio for the asset type. Prefer output_size (exact WIDTHxHEIGHT pixels) when the brief implies physical dimensions; otherwise use aspect_ratio (W:H). GPT Image 2 rules: long edge max 3840px, aspect ratio max 3:1, both dimensions multiples of 16, total pixels 655360-8294400. Presets like 1:1, 3:4, 2:1, 16:9, 9:16 work; custom ratios and pixel sizes are supported. Set output_tier to "4k" for print-ready PNG at creation time (3840px long edge or your exact output_size); default "2k" draft upscales on approve_design. Asset design never enters ad copy: approve_design completes the job. Brand Knowledge photos are not attached for asset design. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_brand

ChatGPT

Use this when the user has no Adly brand yet and wants to set up their business (for example "set up my shop in Adly"). Creates the brand profile Adly designs with: name, category, description, language, colors, fonts, contact details and which contact details show on designs. Requires business_name, category, description and default_language. Do not use to change an existing brand: use update_brand.

create_brand_template

ChatGPT

Create a reusable brand design template. Pass image_url (or upload_product_image first) to save an existing layout immediately. Or pass brief to generate a template candidate job (same as Adly "Generate template"); then poll get_job and call save_brand_template with that job_id (or approve_design) to persist it. Returns template and/or job_id. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_carousel_tips

ChatGPT

Use this when the user wants a carousel, slides or a multi-image post for their business or brand (several swipeable slides for Instagram, LinkedIn or Facebook), designed on their brand kit and ready to publish: tips, a how-to, how to spot the original, reasons to choose the business, a product guide or myths versus facts (for example "make a 5-slide carousel with tips for choosing a running shoe" or "اعمل لي كاروسيل عن طريقة تمييز العود الأصلي"). You write every slide before calling: pass the exact cover and tip slides in carousel_slides. Adly designs the slides you provide; it does not invent slide copy from a vague prompt. Optionally include Brand Knowledge photos via brand_asset_prefs or include_brand_assets. Contact-on-design checkboxes default from the brand profile; pass contact_visibility to override for this job. Returns a job_id; call get_job to check status and media. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_from_image

ChatGPT

Use this when the user has a photo and wants a new post or ad designed from it (if they want to post the photo exactly as it is, use create_ad_copy instead): a product photo to turn into an ad, several product photos for one bundle or collection post, or an old design or an example they like to remake on their brand (for example "turn this product photo into an ad", "make a post like this one" or "حوّل صورة المنتج هذه إلى إعلان"). Upload photos shared in the chat with upload_product_image first. If there is no photo, use create_general_design. Pass image_url for a single base photo, or image_urls (up to 12) to compose one design from several real product photos: a bundle post, a package cover, a set laid out in one frame. With several photos, set use_all_assets so every one is used rather than only anchoring to the first. Optionally reuse a saved brand design template via template_id or template_name (see list_templates). Optionally include Brand Knowledge product photos via brand_asset_prefs or include_brand_assets (place photos are not mixed when a base image is supplied; Brand Knowledge is suppressed when a brand template is applied). Contact-on-design checkboxes default from the brand profile; pass contact_visibility to override for this job. Returns a job_id; call get_job to check status and media. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_general_design

ChatGPT

Use this when the user wants a social media post, ad, graphic or image for their business or brand (to post on Instagram, Facebook, TikTok, LinkedIn or X, or to use as an ad) and has no photo to build it around. Adly makes it on their own brand kit (logo, colors, fonts, language) and it can then be published or scheduled to their connected accounts. Typical requests: an offer or discount, a new product or service, an occasion or holiday greeting, an opening or event, a hiring post, a brand introduction, or each post of a weekly content plan (for example "make an Instagram post for our weekend sale, 20% off everything" or "صمم لي إعلاناً لعرض نهاية الأسبوع"). Adly writes the caption after approve_design. Put the offer, prices, dates and any exact text in the brief, and never add a price, discount or claim the user did not give. For a design built around the user's own photo or an example they want remade, use create_from_image; for several slides, create_carousel_tips; for a logo, flyer or menu, create_asset_design. Optionally reuse a saved brand template via template_name or template_id (see list_templates). To use a real Brand Knowledge storefront/product the same way as the create UI, pass include_brand_assets or brand_knowledge_ids from list_brand_knowledge. Those photos are locked automatically, no special token required. Contact-on-design checkboxes default from the brand profile; pass contact_visibility to override for this job. Returns a job_id; call get_job to check status and media. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_image_to_video

ChatGPT

Backward-compatible shortcut that animates one existing image. For prompt/reference creation and full video options, use create_video. For multi-scene ads, use create_video_scenes. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_landing_page

ChatGPT

Use this when the user wants a landing page, a one-page site or an order page for a product, offer or service, to send ad or bio-link traffic to. Infer the best landing-page brief from the conversation and pass a concise brief (400 chars or fewer). Supports no-product/service pages, one product, or up to 80 products. For product pages, extract each product into products[] with its own description, price, compare price, variants/options, image_url, and optional category_id or category_name when available. Use upload_product_image first for local product images. Optionally pass product_categories[] to predefine saved brand categories. Delivery/shipping is shared across all products. Set act_as_shop=true on product pages when visitors should pick products and add to cart before submitting the lead form or WhatsApp message. Supports capture_mode=form for lead-form pages and capture_mode=whatsapp for no-form WhatsApp-only pages. Returns landing_page_id; call get_landing_page to poll status and get the public URL.

create_multi_mode

ChatGPT

Create Adly multi-product designs from product image/name/price rows. Optionally reuse a saved brand design template via template_id or template_name (see list_templates) — same as Adly multi-mode "from template". Brand Knowledge assets are not used in multi mode. Contact-on-design checkboxes default from the brand profile; pass contact_visibility to override for this job. Returns a job_id; call get_job to check status and media.

create_music_video

ChatGPT

Add Adly-generated background music to your own clips (the Add Music job). Upload each clip with upload_video first. Adly makes music options per clip; poll get_job until music.items[].status is ready, then call approve_music to choose a track. Uses 1 video from the usage limit per clip. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_product_studio_video

ChatGPT

Create an Adly Product Studio video using the active frontend Product Studio flow: preset, product image, optional actor image, optional scene image, aspect ratio, duration, resolution, and style hint. Returns a job_id; call get_job to check status and media.

create_talking_character

ChatGPT

Create a saved cartoon/talking character from a source image, matching the active frontend character creation flow. Generates the character scene image and stores it for the brand.

create_talking_character_video

ChatGPT

Create an Adly cartoon/talking-character video from saved characters, matching the active frontend Talking Characters flow. Each scene chooses a character and one 95-character spoken line.

create_ugc_video

ChatGPT

Use this when the user wants a UGC-style video ad: an AI presenter talks to the camera about their product or service, as a selfie talk, a POV clip, an expert explainer or a voiceover over b-roll (for example "make a UGC-style video ad presenting our new perfume"). The presenter is AI-generated: never present the video as a real customer review or testimonial. Follows the Adly Videos wizard: idea (brief), per-clip spoken lines, actor, optional product, optional Brand Knowledge place photo, optional brand logo. Use list_brand_assets (category storefront/interior) for location, get_brand.custom_actors for brand actors, upload_product_image for product/actor images. Returns a job_id; call get_job to check status and media. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_video

ChatGPT

Use this when the user wants a short video ad or reel (about 4 to 10 seconds; for longer use create_video_scenes): animate a design or product photo, or create a new clip from an idea, from product, model and scene photos, or from an existing video (for example "make an 8 second video ad for my product", "animate this post" or "أريد فيديو قصيراً لمنتجي"). For a person talking to the camera about the product, use create_ugc_video; for a multi-scene commercial up to about 60 seconds, use create_video_scenes. Modes: animate an existing still, or create_new via text_prompt / from_image / reference_images (product + actor/model + scene) / from_video. Brand Knowledge works like the UI picker: list_brand_knowledge then brand_knowledge_ids. Those real photos attach as labeled references (storefront, product, ...) so the brief can name them in plain language and use them anywhere in the video, not automatically as the opening frame. Pass start_frame_url / from_image only when the video should open on a still. Do not invent a storefront or product. Use product_image_url, actor_image_url, scene_image_url for named reference slots. If animate vs create_new is ambiguous, ask once before calling. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

create_video_scenes

ChatGPT

Use this when the user wants a longer commercial with several scenes (about 12 to 60 seconds), like a brand film or a story ad. For one short clip use create_video. Creates a Scene Studio multi-scene commercial (job_type video_scenes). Brand Knowledge works like the UI picker: call list_brand_knowledge (any tag), then pass brand_knowledge_ids. Those photos become @image1...@image3 on every scene. include_brand_knowledge=true fills leftover slots from labeled Brand Knowledge. Recommended: short idea + ai_write=true. Then AI plans 2-6 scenes toward target length (12-60s) and generates scene 1. brief required unless ai_write=true and ai_write_mode=invent. It creates a new draft for the user to review and spends part of the user's Adly usage limit, which cannot be given back once used. Nothing is published until the user asks.

cut_video_clip

ChatGPT

Cut/split a clip at a source time (seconds). Same as UI Cut at playhead.

cut_video_export_music_clip

ChatGPT

Split an export music clip at a timeline time.

delete_brand

ChatGPT

Permanently delete an Adly brand and everything in it. Use only when the user explicitly asks, and confirm with them first. From ChatGPT, confirm_brand_name must be the exact brand name. Uses the MCP default brand when brand_id is omitted.

delete_brand_asset

ChatGPT

Delete one Brand Knowledge asset by asset_id from list_brand_assets.

delete_brand_style_example

ChatGPT

Delete one uploaded Teach Adly your brand style example image by example_id from get_brand_style_learning.

delete_brand_template

ChatGPT

Archive (soft-delete) a saved brand design template. It will no longer appear in list_templates.

delete_landing_product_category

ChatGPT

Delete a saved product category for a brand. Products using it become uncategorized.

delete_talking_character

ChatGPT

Delete a saved cartoon/talking character from the selected brand.

delete_video_clip

ChatGPT

Delete a clip/scene from the Scene Studio timeline. Same as UI Delete.

delete_video_export_music_clip

ChatGPT

Delete an export music clip from the music timeline.

delete_video_layer

ChatGPT

Delete an overlay layer.

delete_video_scene

ChatGPT

Delete a scene via the scenes API. Prefer delete_video_clip for timeline editor parity.

edit_brand_template

ChatGPT

Start or continue an edit session for a saved brand template, then optionally queue an AI design edit. Returns job_id for the edit session. Pass edit_text to request changes (same as Adly Request changes). Poll get_job / get_brand_template_edit_state, then save_brand_template to persist.

edit_talking_character

ChatGPT

Edit a saved cartoon/talking character name, style, or source image. Regenerates the scene image when requested or when the source image changes.

export_video_scenes

ChatGPT

Export/join the Scene Studio timeline into a final MP4 (applies trims, play_speed, transitions, overlays, and mixes the export music bed). Same as UI Export video. Returns export_url / reel URL.

generate_speech

ChatGPT

Turn text into a spoken audio file. Required: text. Optional: voice, quality, language_code, speed, stability, similarity_boost, style, timestamps, previous_text, next_text. Call list_speech_voices first to choose a voice id. Each successful generation uses 0.5 of the video usage limit. Returns audio_url and duration_sec.

generate_video_export_music

ChatGPT

Generate export music bed options for the whole Scene Studio video. Same as UI Generate music options.

generate_video_scene

ChatGPT

Generate one Scene Studio scene (Omni). Uses the video usage limit. Same as UI Generate.

get_brand

ChatGPT

Get the full Adly brand profile: identity, contact info, fonts/colors, contact visibility flags, caption defaults, custom actors (for create_ugc_video selected_actor_id as custom_actor:{id}), WhatsApp bot knowledge profile, and current prompt_instructions from Teach Adly your brand.

get_brand_style_learning

ChatGPT

Get Teach Adly your brand style status: uploaded example images, analysis state, generated instructions, and whether the knowledge has been applied to the brand.

get_brand_template_edit_state

ChatGPT

Get the current edit-session job for a saved brand template (if any), including design media URLs.

get_job

ChatGPT

Use this to follow a design or video you started and to get its media, ad copy and original brief. Designs and videos take one to a few minutes: pass wait_seconds (up to 40) to wait inside this call until the job leaves the generating stage, and repeat until it is ready without asking the user. Returns the status, current stage, media URLs, ad copy, and the original brief/script (user_message, brief, script, script_segments). Use list_jobs when you do not already have the job_id.

get_landing_page

ChatGPT

Get an Adly landing page status after create_landing_page. Poll until status is published or failed. Returns the public URL when ready.

get_meta_ad_account_overview

ChatGPT

Use this when the user asks how their ads are doing (spend, results, reach, whether ads are running) or whether their Meta ad account is connected. Reads the selected brand Meta ad account overview: connection status, currency, account details, and account-level insights. Read-only: it does not create or change campaigns and never exposes tokens.

get_meta_campaign

ChatGPT

Read one Meta campaign, its ad sets, ads, creative thumbnails, and campaign insights. The campaign must belong to the selected brand ad account.

get_video_scenes_state

ChatGPT

Get full Scene Studio timeline state: clips/scenes, trims, play_speed, transitions, overlays, omni options, export music bed, export_url, total duration. Prefer this after create_video_scenes and before any edit.

list_brand_assets

ChatGPT

Alias of list_brand_knowledge. List Brand Knowledge photos (any tag). For Scene Studio pass asset_id values as brand_knowledge_ids on create_video_scenes. For UGC place lock only, filter storefront/interior and pass ugc_scene_place_asset_id.

list_brand_knowledge

ChatGPT

List Brand Knowledge photos for this brand (same as the Brand Knowledge page / Scene Studio Brand Knowledge picker). Every labeled photo is eligible: the tag can be storefront, product, service, project, or any other Brand Knowledge category. Returns asset_id, category, caption, image_url. For Scene Studio: pass chosen asset_id values as brand_knowledge_ids on create_video_scenes (first id = @image1). Alias of list_brand_assets.

list_brands

ChatGPT

Use this first, before creating anything, to find which Adly brand (business) to work on. Search by name, confirm the right match with the user, then call select_brand with its brand_id. If the user has exactly one brand, use it without asking. Without a search term only a small recent sample is returned, never the full list.

list_jobs

ChatGPT

List a brand's jobs, including completed work. Defaults to completed. Returns each job's original brief or script so you can find a job without guessing IDs. Use list_pending_jobs for jobs that are still waiting on the user and list_published_posts for posts that went out.

list_landing_product_categories

ChatGPT

List saved product categories for a brand. Use category_id or category_name on landing products to group catalog sections on generated store pages.

list_meta_ads

ChatGPT

Read Meta ads in the selected brand ad account. Can filter by campaign_id, adset_id, status, and effective_status, with creative metadata and optional insights.

list_meta_adsets

ChatGPT

Read Meta ad sets in the selected brand ad account. Can filter by campaign_id, campaign objective, campaign buying_type, optimization goal, status, and effective_status.

list_meta_campaigns

ChatGPT

Use this when the user asks about their Facebook and Instagram ad campaigns: which are running, how each one performs, or to find a campaign. Reads Meta campaigns in the selected brand ad account with filters for objective, buying_type, status, effective_status, and optional insights. Read-only.

list_pending_jobs

ChatGPT

List open MCP jobs (including publish-ready). blocks_new_creates=true means the job holds a creation slot. Use cancel_job to close stuck jobs. create_* auto-discards oldest slot-holders when the cap is reached. Does not return completed jobs; use list_jobs for finished brand work and original briefs/scripts.

list_published_posts

ChatGPT

List organic posts already published from Adly to Facebook, Instagram, LinkedIn, X or TikTok for the selected brand. Returns stored post ids, Instagram permalink / LinkedIn / X / TikTok URL when available, ad copy, and generated media. Read-only; does not publish or call the networks live. Scheduled posts are in list_scheduled_posts; unpublished completed jobs are in list_jobs.

list_scheduled_posts

ChatGPT

Use this to see a brand's upcoming scheduled posts, or to check whether a publish_now / schedule_post went live (pass job_id). Returns status, time, networks, caption preview, live links once published, and the error if one failed. Read-only.

list_social_accounts

ChatGPT

Use this when you need to know where a brand can post. Returns the Facebook Page, Instagram, LinkedIn, X and TikTok connection status (connected, expired, disconnected), whether the plan allows X, and a connect_url for fixing missing ones. Read-only.

list_speech_voices

ChatGPT

List every available speech voice, quality, language code, and generation parameter for generate_speech. Call this before generate_speech so you can pick a voice id. Does not use the usage limit.

list_talking_characters

ChatGPT

List saved cartoon/talking characters for the selected brand. Use the returned character ids with create_talking_character_video, edit_talking_character, clone_talking_character, or delete_talking_character.

list_templates

ChatGPT

List saved brand design templates for the selected brand. Use template_name or template_id when creating announcement, before_after, idea_swipe, from_image, or multi_mode jobs from a template. Also use with create_brand_template / edit_brand_template / rename_brand_template / delete_brand_template / save_brand_template.

merge_video_scenes

ChatGPT

Merge two adjacent ready scenes into one. Same as UI merge.

open_content_board

ChatGPT

Use this when the user wants to see their Adly content at a glance: the latest designs and videos of a brand with their status, and the posts scheduled to go out. Opens the Adly content board. It only reads. Works with no arguments (uses the selected brand).

plan_video_scenes

ChatGPT

AI-plan 2 to 6 Scene Studio scenes from the master brief (Scene Studio AI plan). Merges around already-ready scenes. Same as the UI AI plan flow.

poll_video_export_music

ChatGPT

Poll export music generation status until ready/failed.

poll_video_scenes

ChatGPT

Poll in-flight Scene Studio generation; finalize ready scenes / heal stuck tasks. Call while status is generating.

process_job

ChatGPT

Queue processing for an existing Adly job and return immediately. Use get_job to poll status and media.

promote_video_scene_take

ChatGPT

Promote an alternate take to the main scene video. Same as UI Use as main.

publish_now

ChatGPT

Use this when the user wants a finished Adly design, carousel or video posted right away to Facebook, Instagram, LinkedIn, X and/or TikTok. Queues the post; it goes live within about a minute (check with list_scheduled_posts). Uses the job's approved ad copy as the caption unless caption is given. Confirm the networks and caption with the user before calling. Do not use for future times: use schedule_post.

regenerate_video_scene

ChatGPT

Regenerate one Scene Studio scene and mark later scenes stale for continuity. Same as UI Regenerate.

rename_brand_template

ChatGPT

Rename a saved brand design template.

reorder_brand_assets

ChatGPT

Reorder Brand Knowledge assets within one category. Pass the full ordered list of asset ids for that category.

reorder_video_timeline

ChatGPT

Reorder Scene Studio clips. Pass timeline as an ordered array of clip ids. Same as drag-reorder in the UI.

request_copy_edit

ChatGPT

Use this when the user wants the caption changed: shorter, another tone, add the price, phone number or a link, or another language. Queues the copy edit; poll get_job for the new caption.

request_design_edit

ChatGPT

Use this when the user wants changes to a design Adly made (for example "make the price bigger", "change the background to white" or "fix the spelling"). Queues the edit; poll get_job for the new version. For multi_mode after product variants exist, pass product_id (or product_ids / apply_to_all_products) from get_job media[].product_id. Before the template is applied to products, omit product_id to edit the master template.

restore_video_timeline

ChatGPT

Restore a full Scene Studio project snapshot (undo/redo). Pass the project object previously returned by get_video_scenes_state.

save_brand_template

ChatGPT

Persist a brand template. For a saved-template edit session: pass template_id + job_id (from edit_brand_template). For a newly generated candidate job: pass job_id from create_brand_template (generate path), the same as approving/saving as template. Optional name renames on save.

save_landing_product_category

ChatGPT

Create or rename a saved product category for a brand. Returns the category id to use on landing products.

schedule_post

ChatGPT

Use this when the user wants a finished Adly design, carousel or video posted later to Facebook, Instagram, LinkedIn, X and/or TikTok. Scheduling the same job again moves its time. Confirm the time, networks and caption with the user before calling. Do not use to post now: use publish_now.

select_brand

ChatGPT

Use this after list_brands to set the brand (business) that the next designs, videos and posts are made for. Get the brand_id from list_brands first.

select_video_export_music

ChatGPT

Select an export music track, set volumes, or clear the music bed. Same as UI music select/mix.

set_landing_page_act_as_shop

ChatGPT

Enable or disable Act as shop on an existing product landing page. When enabled, visitors can pick products, add to cart, and submit order details through the lead form or WhatsApp. Requires a product page (has_product). For best cart UX on pages generated without shop mode, regenerate after enabling.

set_landing_page_capture_mode

ChatGPT

Switch an existing Adly landing page between lead-form capture and WhatsApp-only capture. Use capture_mode=form to restore a lead form; if no form exists, Adly uses the default form unless form_template_id is supplied. Use capture_mode=whatsapp to replace the lead form with a large WhatsApp CTA; requires whatsapp_phone.

show_job

ChatGPT

Use this to show the user a job's designs or video visually in the chat (an Adly preview card with the media, caption and status). Call it once when get_job shows the media is ready, or when the user asks to see a job. It only reads; it never changes the job.

update_brand

ChatGPT

Update an existing Adly brand. Only fields you pass are changed; omitted fields keep their current values. Supports contact visibility, caption defaults, logo URL, custom actors, and WhatsApp bot profile.

update_brand_asset

ChatGPT

Update a Brand Knowledge asset category and/or caption (label photos after upload). Categories: storefront, interior, product, service, project, before_after, design_pattern, team, packaging, menu_or_price, food_dish, vehicle, equipment, property_unit, or uncategorized.

update_landing_page_products

ChatGPT

Update products, categories, shipping, or shop mode on an existing classic landing page without regenerating the whole page. Mirrors the Stores and landing pages manage-products UI.

update_video_clip

ChatGPT

Update one Scene Studio clip/scene settings: title, prompt, spoken_script, onscreen_text, duration_sec, trim_in/trim_out, play_speed (0.5 to 2), transition_out, transition_out_sec. Same as the clip inspector.

update_video_export_music_clip

ChatGPT

Update an export music clip (timing, fades, volume).

update_video_layer

ChatGPT

Update an overlay layer (position, timing, text, style, image).

update_video_scene

ChatGPT

Update a scene via the scenes API (prompt/title/spoken/onscreen/duration/trim). Prefer update_video_clip for timeline editor parity including play_speed and transitions.

update_video_scenes_options

ChatGPT

Update Scene Studio video options (voice narration, on-screen text, music mode, narrator mood/sex/age/accent, export volumes). Same as the omni options drawer.

upload_brand_assets

ChatGPT

Alias of upload_brand_knowledge. Upload Brand Knowledge photos (same as Brand Knowledge UI). Tag/category optional; label with update_brand_asset. Max 10 images per call.

upload_brand_knowledge

ChatGPT

Upload Brand Knowledge photos (same as the Brand Knowledge page). Pass HTTPS URLs, existing Adly /uploads/... paths, and/or base64 images. Tag/category is optional on upload (defaults to uncategorized); label with update_brand_asset. Then use the returned asset_id with create_video_scenes brand_knowledge_ids. Max 10 images per call. Alias of upload_brand_assets.

upload_brand_style_examples

ChatGPT

Upload 1-10 brand style example images for Teach Adly your brand. Pass HTTPS URLs, existing /uploads/... paths, and/or base64-encoded images. Upload at least 3 examples before calling analyze_brand_style. Uses the same brand-strength example store as the Adly UI.

upload_product_image

ChatGPT

Use this when the user shares a photo in the chat (a product, their shop, an old design or an example they like) and a create tool needs it. Uploads the image to Adly and returns an image_url for create_from_image, create_multi_mode, create_ugc_video, create_product_studio_video, create_talking_character, create_landing_page product images, create_asset_design reference directions, create_video, create_image_to_video, or create_video_scenes. Pass base64-encoded image bytes (raw base64 or a data URI such as data:image/jpeg;base64,...).

upload_video

ChatGPT

Upload a video (MP4, MOV, WEBM, M4V; up to 100 MB) to the brand and return a video_url for create_music_video or create_ad_copy. Small files (under about 25 MB): pass video_base64. Public files: pass video_url (https). Larger files: send chunk_base64 pieces of at most 20 MB each, in order with the same upload_id (chunk_index 0, 1, 2 ...) and final=true on the last piece; only the final call returns video_url.