account_summary
ChatGPTA short overview of the account (for an admin: of the selected client): posts waiting for review, scheduled and published, the next publication, connected channels, open tickets, unread notifications, whether analytics is connected and which reports are on.
ai_visibility_check
ChatGPTThe free check of the IN-STORE blog ("Does AI see your site?"): looks at a public website the way ChatGPT, Claude and Perplexity do (answer to AI crawlers, robots.txt per crawler, text without JavaScript, title and description, structured data, sitemap, llms.txt) and returns a score with each check and what to fix. url: the site address. A few checks per ten minutes.
analytics_ask
ChatGPTAnswers a question about the account's website and social analytics in plain words, through the cabinet's own Performance assistant (same daily limit). Spend and other money figures are removed from the answer: they are available only in the web cabinet.
analytics_ask_history_clear
ChatGPTThe trash button of the Performance assistant: removes every question the account asked it that the server remembers. The account background written by the team stays. Without confirm it returns a preview; with confirm: true it makes the change.
analytics_report
ChatGPTThe cabinet Performance pages as data, without any money figure. what: traffic (Google Analytics), search (Search Console), advertising (ads results without spend), drill (one view broken down: view and dimension, of = the parent value), keywords (search terms worth writing about), paid_organic (paid vs organic search terms), anomalies (unusual changes), content_search (which pages bring search visitors), status (what is connected). For a question in plain words use analytics_ask.
analytics_report_send
ChatGPTEmails the cabinet's performance report (PDF) exactly like "Email report" on the Performance page: platforms (traffic, search, advertising), the period, optional sections and recipients (up to five; default the account email). The report itself may contain money figures; it goes to the recipients, not to this conversation. Without confirm it returns a preview; with confirm: true it makes the change.
brand_file_add
ChatGPTAdds a file to the account from a public https address (the server downloads it; nothing is published). what: overlay (the brand frame of one picture shape: ratio 1x1, 4x5, 16x9 or 9x16, fileUrl of a PNG; replaces the frame of that shape), carousel_logo (a logo slides can use: fileUrl, fileName optional), video (a video for the video library: fileUrl, name optional; MP4, WebM or MOV), media_folder (admin only: create the client's media folder on Drive). Without confirm it returns a preview; with confirm: true it makes the change.
brand_file_remove
ChatGPTRemoves the brand frame of one picture shape (what: overlay, ratio 1x1, 4x5, 16x9 or 9x16); its file goes to the Drive trash. Pictures made before keep their frame. Without confirm it returns a preview; with confirm: true it makes the change.
brand_system_run
ChatGPTbuild: creates the brand, frames and logos folders this account's pictures need (safe to repeat, creates nothing twice). test: prints a test picture with this account's own template on a blank square of ours, to prove the template works; nothing of the account is touched. Without confirm it returns a preview; with confirm: true it makes the change.
carousel_get
ChatGPTwhat: job (the carousel of the post: status, formats, slide pictures), content (the slide words), revisions (saved versions), changes (what changed in one version: rev), photos (library photos a carousel may use).
carousel_manage
ChatGPTThe carousel buttons of the cabinet. action: start (make a carousel for the post: formats, theme, slides, style, focus, cta, photos, force to make a new one), use (make the slides the post's pictures: jobId, format), unuse (go back to the post's own picture), copies (save the slides into the library: jobId, format, title), edit_slide (one slide: jobId, format, at, field, value, rev; allFormats), edit_slides (several: jobId, format, rev, changes), reorder (jobId, format, rev, plan = slide numbers in the new order), rollback (jobId, format, rev). Read carousel_get first for jobId and rev. Without confirm it returns a preview; with confirm: true it makes the change.
carousel_slide_preview
ChatGPTSlide Studio preview: renders one slide of a post's carousel with changes that are not saved, and saves nothing. postId or rowNumber (media_plan_list), jobId (carousel_get), format (a slide shape such as portrait45), at (the slide, from 0); optional fields ({ field: text }), restore (fields back to the shared text), frames ({ field: {x,y,w,h} }). Returns the layer plan and whether the slide fits; the slide HTML is only measured. carousel_manage edit_slide saves a change.
client_document_manage
ChatGPTThe account's own documents (kept as text for 30 days). action: add (fileUrls: 1 to 5 public https addresses of PDF, Word, text or HTML files; returns each document's id, text, a suggestion whether it is a finished post or notes, and the post it would make), make_post (documentId, channels optional: makes a post for review from a finished-post document, with a new picture; follow it with media_assets_get document_job). Notes for the writer go to idea_manage create documentIds. Without confirm it returns a preview; with confirm: true it makes the change.
companies_list
ChatGPTThe companies the signed-in person may work in: their own company first, then every company that gave them access, each with its id, name and the person's role there (owner, approver or viewer; a viewer reads only), and which one this connection works in now. company_switch changes it.
company_switch
ChatGPTMakes this connection work in another company the person has access to, like "Switch company" in the cabinet: companyId from companies_list; the person's own company id goes back home. From then on every tool reads and changes that company, with the person's role there (a viewer reads only) and the cabinet sections that company has. Billing stays with the person and is not offered while working in another company. The access is re-checked on every call; when it ends, the connection is back in its own company.
competitors_manage
ChatGPTaction: list (add one competitor site: add, or remove one: remove), pages (a competitor's listing or YouTube page: host, listingUrl, youtubeUrl), toggle (follow a competitor or not: host, enabled; replace to swap), refresh (read competitors again now: hosts, networks; paid networks count against the account limit). Without confirm it returns a preview; with confirm: true it makes the change.
competitors_read
ChatGPTThe Competitors tab: the competitors of the account, their pages and what was found on their sites and social networks.
connection_status
ChatGPTWhich outside accounts are connected for analytics and publishing (Google Analytics, Search Console, Google Ads; for an admin also X). Connecting one is a sign-in the person does on the provider's page, so it cannot be done from here: the answer gives the cabinet address where they press Connect.
connections_get
ChatGPTThe Connections overview of the cabinet: every channel of the account (social networks, the website, Google Business Profile, X, Google Analytics and Search Console, Google Ads) with whether it is connected and working. Connecting one is done in the cabinet; connection_status gives its address.
digest_schedule_get
ChatGPTShows the Journalist digest schedule (which days topic digests arrive) and the service categories the topics are chosen for.
digests_manage
ChatGPTaction: request (ask the Journalist for a digest now: topic, format, referenceLinks), start (write a post from a digest topic: sessionId, topicNo), dismiss (drop a topic: sessionId, topicNo), settings (digest settings: settings as digests_read settings shows them), schedule (the digest schedule: schedule as digest_schedule_get shows it). Without confirm it returns a preview; with confirm: true it makes the change.
digests_read
ChatGPTThe Journalist digests: what: digests (the days and the topics found, with sessionId and topicNo for digests_manage), settings (digest settings).
eblast_file_attach
ChatGPTeBlast tab: attaches a picture (PNG, JPEG, WebP, GIF) or a PDF to the company's eBlast files, as the page's attach button does; the server downloads it from its public https address (8 MB at most). fileUrl; fileName optional (with its extension); blastId optional (the eBlast it belongs to, from eblast_get). Without confirm it returns a preview; with confirm: true it makes the change.
eblast_get
ChatGPTThe eBlast tab: the e-mail campaigns prepared for the account and where each stands. what: overview (every open eBlast with status, version, sending days, what is confirmed, open requests, results counts of past sends, topic requests), blast (one eBlast in full: banners, text blocks, comments, own wording, checklist and the e-mail text; blastId), events (the full journal of changes, admin only; since = the last seq already read), notify (the addresses that get the eBlast notices; empty means the account address). Results are counts; no contact list is shown.
eblast_send
ChatGPTSends e-mail from the eBlast tab, as its two buttons do. action: approve (approve one eBlast: blastId, version; once approved the team's sender e-mails it to the company's contacts on its sending days; the cabinet allows it when banner, text and days are confirmed and the checklist is done), test_copy (a test copy of one eBlast to the signed-in person, cc optional; allowed once banner, text and days are confirmed and the checklist is done). Two confirmations from the person are needed: without confirm it returns a preview; with confirm: true it returns a second question and a one-time confirmCode; with confirm: true and that confirmCode it sends.
eblast_update
ChatGPTOne change on the eBlast tab, as its buttons make it. For one eBlast (blastId and version from eblast_get): banner (choose banner option chosen), confirm (area banner, text or dates as right or not: on), comment (a comment on a block, the banner, dates or general: on, text), remove_comment (the person's own comment: on, index), wording (new wording for a text block: on = block id, text; the original text takes the wording back), add_block (suggest a new text block: text, where), check (a before-release checklist item: item, on), days (new sending days: days as eblast_get blast shows them, with new iso dates; only before the first day is sent), withdraw (an approved eBlast back to review: nothing more is sent until it is approved again). For the account: topic_request (ask for a new eBlast: topic, details, week, links), notify (who gets the eBlast notices: emails, empty for the account address). An approved or sent eBlast changes only by withdraw. Approving and test copies are eblast_send. Without confirm it returns a preview; with confirm: true it makes the change.
fact_check
ChatGPTThe fact-check panel of a post. action: run (check the article against the web and its sources: AI work), confirm (claims the person confirmed are right: claims = claim numbers from published_get fact_check), dismiss (claims to ignore), mark_fixed (claims fixed in the text), rewrite (AI suggests a corrected sentence: issue, sourceSays, instruction optional; nothing is saved). A correction itself is made with post_update field. Without confirm it returns a preview; with confirm: true it makes the change.
fact_check_fix
ChatGPTAfter fact_check run. action: suggest_fix (AI writes a corrected article from the saved verdicts; nothing is saved: put the corrected text on the post with post_update field Post Body (Website)), recheck (re-checks only edited passages: passages = the edited paragraphs as they read now; the other saved verdicts are kept). Without confirm it returns a preview; with confirm: true it makes the change.
fact_check_rewrite_suggest
ChatGPTAn AI suggestion of a corrected sentence for one fact-check finding of a post article, like Rewrite in the cabinet fact-check panel: issue (what is wrong), sourceSays (what the source says, optional), instruction (optional). Nothing is saved: the article is changed with post_update field Post Body (Website). This uses paid AI work and records usage. Without confirm it previews the request; with confirm: true it runs.
file_upload
ChatGPTUploads files into the account's media folder on Drive, as the Upload buttons of the intake form, the Media Library and Media Plan "+ New post" do; the server downloads each file from its public https address. field: logoFiles, brandAssetFiles or imageSampleFiles (pictures only). fileUrls: 1 to 10 addresses (25 MB each). Without confirm it returns a preview; with confirm: true it makes the change.
help_search
ChatGPTSearches the Help Centre of the person's brand (how to connect social networks, the intake, the media plan, tickets, settings) and returns the best matching articles with their text. Use it to answer "how do I..." questions about the platform.
history_get
ChatGPTThe History tab: what was done on the account (by the client, the team and the system), newest first, worded for the client. title or row narrows it to one post; beforeId and limit page through it.
idea_decide
ChatGPTStarts writing an idea from the Ideas list (decision start) or declines it (decision decline), like the cabinet. Use rowNumber and title from ideas_list. Without confirm it returns a preview; with confirm: true it makes the change.
idea_manage
ChatGPTThe buttons of the Ideas tab. action: create (title; mainIdea, notes, research, primaryKeyword, secondaryKeywords, articleLength, referenceLinks, documentIds of documents added with client_document_manage as notes for the writer, autoStart false keeps it in Ideas, force true adds it although a post with a similar title exists), field (rowNumber, field, value), start (rowNumber: the Writer starts writing the article), not_interested (rowNumber), restore (rowNumber), bulk_not_interested and bulk_restore (rows, at most 100). Rows come from ideas_list; with field, start, not_interested and restore, title (optional) is checked against the row and a different title is refused. Without confirm it returns a preview; with confirm: true it makes the change.
idea_submit
ChatGPTAdds a topic idea to the account's Ideas, exactly like Add idea in the cabinet: it is researched first, then written. For an admin it is added for the selected client as a team-added topic. Without confirm it returns a preview; with confirm: true it makes the change.
ideas_list
ChatGPTLists the topic ideas of the account (for an admin: of the selected client): the client's own ideas and the topics the Journalist found, with their status.
image_generate
ChatGPTMakes a new picture exactly like Generate in the cabinet media library: prompt (what should be on it), size (square, portrait, landscape), quality (low, medium, high), overlay (none, logo, full: the account's own template; full also prints barText), enhance (default true). With rowNumber or postId the picture is made for that post (named after it and counted against it). Each picture costs and counts against the account's daily limit; media_read generate_options lists the accepted choices. Without confirm it returns a preview; with confirm: true it starts the picture and returns at once with status started and a jobId. A picture takes about 2 minutes; image_generate_status with that jobId gives the result (the url, which post_update action picture takes as imageUrl). The picture lands in the media library either way.
image_generate_status
ChatGPTThe state of a picture started by image_generate (jobId from its answer). status running: seconds so far and how many seconds to wait before reading again. status done: url (the address post_update action picture takes as imageUrl), pictureUrl, the library file names, and the post it was made for. status failed: the error code and what went wrong. status not_found: unknown here, older than an hour, or the server restarted; the picture may still be in media_library_list.
intake_competitors_get
ChatGPTThe competitor list of the intake answers, as Settings -> Your answers shows it: each competitor with its address and host, notes written without an address, and, when the account has the Competitors feature, whether each one is followed and how many posts and videos were found.
intake_competitors_update
ChatGPTAdds one competitor to the intake answers (add = its web address) or removes one (remove = its host from intake_competitors_get), like the list in Settings -> Your answers. With the Competitors feature an added competitor is read from its site straight away. Without confirm it returns a preview; with confirm: true it makes the change.
intake_draft_get
ChatGPTThe intake questionnaire as the intake page holds it: the answers given so far (field name and value), the names of the fields still empty, the step the person is on (currentStep, 0 = first), the intake version, whether it is submitted, whether a new pass is open (intakeEditing) and how many passes are left. intake_status gives the onboarding stage around it.
intake_draft_save
ChatGPTSaves answers into the intake draft, like typing them on the intake page (its autosave): answers = { field name: value } with the field names intake_draft_get returns (a channel group field takes an object of the channels to change); currentStep moves the page to that step (optional); expectedVersion = the intakeVersion read (optional; a newer intake is refused, never overwritten). Open before the first submit and during a new pass; one answer of a submitted intake is changed with settings_update intake_answer. Without confirm it returns a preview of each change; with confirm: true it makes the change.
intake_prefill_cancel
ChatGPTStops an intake AI research that is still busy, like Stop on the intake page (jobId from intake_prefill_status). A stopped research writes nothing; the step is then filled in by hand. Without confirm it returns a preview; with confirm: true it makes the change.
intake_prefill_start
ChatGPTStarts the AI research that fills the next intake step from the answers so far, like Next on the first three steps of the intake page: fromStep 0, 1 or 2 (the step just answered; default the current step). It runs in the background on the saved draft; intake_prefill_status shows it, intake_prefill_cancel stops it. The answers it finds are written into the next step of the draft. Without confirm it returns a preview; with confirm: true it makes the change.
intake_prefill_status
ChatGPTThe state of the AI research of an intake step: jobId for one job, or none for the job running now. status queued, researching, applying or retry_wait (busy), succeeded (its answers are in the draft), failed or conflict (with the reason), cancelled.
intake_redo
ChatGPTStarts a new pass of a submitted intake, like "Redo the intake" in the cabinet: the draft starts again from a blank form and the intake page opens; the submitted answers stay in force until the new pass is submitted. Each account has a limited number of passes (intake_draft_get intakeRemaining). Without confirm it returns a preview; with confirm: true it makes the change.
intake_status
ChatGPTShows where the intake (the onboarding questionnaire) stands: registered, in progress, submitted, set up, or being redone; quick edits since submission; integration channels connected and pending.
intake_submit
ChatGPTSubmits the intake like Submit on its last step: the answers saved in the draft become the account's intake, one pass is counted, and the cabinet sends its usual e-mails (the onboarding e-mail to the account, a notice to the team). The cabinet refuses an intake with required answers missing and names them. Without confirm it returns a preview; with confirm: true it makes the change.
media_alt_suggest
ChatGPTAn AI suggestion of alt text for one picture (url of a library picture or a picture on a post), exactly like Suggest alt text in the cabinet. This uses paid AI work and records usage. Without confirm it previews the request; with confirm: true it runs. The alt text of a post is changed separately with post_update field Image Alt Text.
media_assets_get
ChatGPTwhat: overlays (the brand frame of each picture shape: 1x1, 4x5, 16x9, 9x16), carousel_logos (the logos slides can use), logo_status (what the Add logo window knows about one picture: source = its address, clean optional), document_job (where making a post from a document stands: jobId from client_document_manage make_post).
media_check
ChatGPTChecks that change nothing. what: inspect (whether a picture or video fits the networks of a post: url, channels; kind, format, bytes, width, height optional), preview (the picture with the account's brand frame drawn on it, as the post would carry it: url, brandFrame default true), image_url (whether an image address can be used: url).
media_delete
ChatGPTDeletes one file from the account's media library, exactly like Delete in the cabinet (fileId from media_library_list). The cabinet refuses a file a post still uses. Without confirm it returns a preview; with confirm: true it makes the change.
media_edit
ChatGPTLibrary work exactly like the cabinet media library. action: rename (fileId, name), tags (fileId, tags), crop (url, rect {x, y, w, h} as fractions 0 to 1; saved as a NEW picture), hide or unhide (url), import_url (url of a picture on the web; returns an address on our server to use in posts), fix_for_channels (url, channels, crop: a new copy that every listed network accepts; the original stays), alt_suggest (url: AI suggests alt text, nothing is saved), print_text (url, text: prints the account template with that text on a NEW copy). Deleting is media_delete. Without confirm it returns a preview; with confirm: true it makes the change.
media_fix
ChatGPTMakes NEW copies that fit the networks; the originals and the post are not changed. what: gallery (urls of up to the post's photos, channels; crop true also cuts a photo whose shape a network refuses), video (url, channels: compresses to the strictest size). Put a copy on the post with post_update or post_media_update. Without confirm it returns a preview; with confirm: true it makes the change.
media_library_list
ChatGPTLists the newest photos and videos in the media library (photos from the client's Drive folder, videos from the platform library). url is where the file lives (for a photo, its Drive page). To put a photo on a post, pass its pictureUrl to post_update (picture, pictures, gallery, channel_media) or post_create (userImageUrl).
media_plan_list
ChatGPTLists posts of the media plan by status: review (waiting for the client's approval), scheduled, published, processing (being written or revised), rejected, deleted (removed from the site), or all. Returns counts per status and, for each post, its rowNumber and postId (use them with post_get), title, schedule, article URL and channels.
media_plan_status
ChatGPTwhat: next_slot (the time "Next free slot" gives when a post is approved now, and what "Publish now" does: rowNumber of the post optional), pulse (a fingerprint of which posts exist and their status; a different fingerprint means the plan changed).
media_read
ChatGPTwhat: generate_options (sizes, qualities and text-on-picture choices image_generate accepts for this account), tags (library tags), hidden (pictures hidden from the library), auto_images (pictures the system made), logos (marked logo pictures), caption (the text printed on a picture: needs url), requirements (what each social network accepts: optional channels).
notifications_list
ChatGPTLists the newest notifications of the signed-in person (the cabinet bell), unread first.
notifications_mark_read
ChatGPTMarks one notification (id from notifications_list) or all unread notifications of the signed-in person as read.
picture_logo_add
ChatGPTThe Add logo window. action: prepare (the picture the logo goes on: source = picture address; clean optional; may make, once, a copy with the picture's words and without the printer's logo), apply (makes a NEW picture with the logo: source, logo = a logo address from media_read logos, rect {x, y, w} as parts of the picture from 0 to 1, format auto, webp, jpeg or png, variant main or light, backing {kind none, shade or plate; strength; color dark or light}). The source picture is not changed; the new address can be put on a post with post_update. Without confirm it returns a preview; with confirm: true it makes the change.
post_action
ChatGPTChanges one post of the media plan exactly like the cabinet buttons: approve (optionally for a time, otherwise the next free slot), reschedule (needs scheduledAt), reject, request_changes (needs a comment for the writer), back_to_review, restore. Only the actions the cabinet offers for the post's current status are accepted. Without confirm it returns a preview; with confirm: true it makes the change.
post_create
ChatGPTAdds a new post to the media plan exactly like Create post in the cabinet. title and bodySocial are required; the website text, call to action, hashtags (at most five are kept), alt text, picture address, channels and time are optional. status pending_approval (default) puts it on the review list; scheduled publishes it at scheduledAt (Toronto time, YYYY-MM-DD HH:MM). If a post on the same topic exists the cabinet says so and nothing is created; repeat with force: true to create it anyway. Without confirm it returns a preview; with confirm: true it makes the change.
post_get
ChatGPTReturns one post of the media plan in full: status, schedule, the social text and call to action, the website article (cut to a readable length), SEO title and description, channels and their published links, and the fact-check result. Identify the post by rowNumber or postId from media_plan_list.
post_media_update
ChatGPTThe picture block of a post. intent: promote_photo (make one of the post's photos the website picture: url), replace_photo (index from 1, url), use_video (url of a video, or from: article for the article's own video), remove_video, stop_slides (stop using the slides), restore_state (bring back a recorded state from post_pictures: restoreOf and its main, photos, video). Addresses from media_library_list or image_generate. A published post's website copy follows. Without confirm it returns a preview; with confirm: true it makes the change.
post_pictures
ChatGPTThe pictures a post has had, newest first, with who changed them and when (the cabinet's picture history). Use a picture address from here with post_update action picture to bring an earlier one back.
post_publish_action
ChatGPTThe post buttons that post_action does not cover. action: publish_now (publish to the post's networks now, or right after the post just before it), retry_publish (publish again a scheduled post the publisher flagged with an error), delete (hide the post from the plan for good; nothing is deleted from networks). publish_now and retry_publish need two confirmations: without confirm a preview, with confirm: true a second question and a one-time confirmCode, with confirm: true and that confirmCode the change. delete: without confirm a preview, with confirm: true the change.
post_update
ChatGPTChanges one post through the cabinet route of the same button. action: field (one text field: field and value; fields: Post Title, Main idea, Post Body (Social Media), Call To Action, Hashtags, Post Body (Website), Call To Action (Website), Website META Description, Website SEO Title, Website Slug, Primary Keyword, Secondary Keywords, Image Alt Text), picture (the main picture: imageUrl = the pictureUrl from media_library_list or the url from image_generate; a Drive link of a library file is turned into that form; empty string removes it), pictures (add photos: imageUrls), gallery (the social album: mainUrl and galleryUrls, mode as the cabinet uses it), channels (where it goes: channels), channel_words (own words for one channel: channel and entry), channel_media (own pictures or video for one channel: channel and media { photos: [addresses], video: address }), recaption (print the post text on the main picture again: barText optional, logo optional), recaption_photo (the same on one photo: photoUrl or photoIndex), article_link (admin only: value). A published post: a change of its text or picture also reaches the website copy, as in the cabinet. Without confirm it returns a preview with the current value; with confirm: true it makes the change.
post_x_link_set
ChatGPTThe X link switch of one post: on true adds the article link to the post on X (X charges more for a post with a link), on false leaves it out. Without confirm it returns a preview; with confirm: true it makes the change.
published_get
ChatGPTwhat: channels (where one post went and its links on each network: rowNumber), redirects (website address redirects state), video (the video of one post: postId), video_library (the account's videos), fact_check (saved fact-check results of all posts), sources (source material being turned into ideas or articles).
published_manage
ChatGPTThe buttons of a published post. action: unpublish (take the article off the website), restore (put a removed article back), push_field (send one field's current plan value to the live article: field, e.g. Website SEO Title; Website Slug also registers a redirect), delete_from_socials (delete the post from channels: facebook, linkedin, telegram, instagram, x). These change what the public sees. Without confirm it returns a preview; with confirm: true it makes the change.
reports_manage
ChatGPTaction: settings (change the Reports page settings: settings = the fields exactly as reports_read status shows them, only the ones to change), send_now (send a report now: kind, period). Without confirm it returns a preview; with confirm: true it makes the change.
reports_read
ChatGPTThe Reports page: what: status (which reports are on, recipients, next run, settings), runs (past sends), preview (get the cabinet link to build a sample there; rendered reports can contain money). Use the kind returned by status.
reports_status
ChatGPTShows which performance reports (weekly, monthly) are switched on, when the next one runs and how the last one went.
settings_get
ChatGPTReads one part of the settings the system works with for this account, exactly as the cabinet stores them: channels, channel_registry, integration_channels, styles, carousel_templates, brand_system, link_condom, intake_answers (the intake answers the content, pictures and research are built from), brand_voice, text_guidelines, image_config, default_channels, notifications, social_article_link, auto_reject (admin). Read before suggesting a change with settings_update.
settings_update
ChatGPTChanges one setting of this account through the same server route the cabinet uses. what: channel (turn a channel on or off, or set its page id, address or note; needs channel), style (needs variable and value; names from settings_get styles), link_protection (enabled true or false; switches the outbound-link protection on the client website), intake_answer (needs field and value; names from settings_get intake_answers; subKey for one channel of a channel group), brand_voice (voice), text_guidelines (value), image_config (values), default_channels (channels), notifications (enabled, recipients), social_article_link (enabled), preferences (autoRejectEnabled, autoRejectDays, autoRejectIdeasEnabled; a client also ticketNotificationsEnabled, dictationEnabled), logo_mark (url of a library picture, enabled = is a logo), logo_light (url, lightUrl), overlay_url (admin only: ratio, url of the frame). Without confirm it returns a preview with the current value and what will change; with confirm: true it makes the change.
social_report_get
ChatGPTThe Social media tab of Performance as data: view facebook (Page), instagram, x or google (Business Profile). Totals with the previous period, a daily series, every post in the period marked as published by our automation (with its Media Plan post) or other, and our published posts not seen on the account. Period as preset or from and to. No money figures.
source_ingest
ChatGPTSource material like the cabinet "From a source" form. action: start (url of a page, and/or fileUrls and referenceLinks; target idea (default) or article), summary (a short summary of one url, nothing is saved). Follow the work with published_get sources. Without confirm it returns a preview; with confirm: true it makes the change.
source_summary
ChatGPTA short AI summary of one web page (url), like the summary in the cabinet "From a source" form. Nothing is saved; an idea or an article is made from a source with source_ingest. This uses paid AI work and records usage. Without confirm it previews the request; with confirm: true it runs.
story_get
ChatGPTwhat: story (the post's story versions and which one is chosen), text (the words on the chosen story).
story_manage
ChatGPTThe story buttons of the cabinet. action: want (wanted true or false: whether the post gets a story), make (draw a story: fromPost true for the post picture, or jobId of a carousel), select (versionId), rollback (rev), edit_text (field, value). Read story_get first. Without confirm it returns a preview; with confirm: true it makes the change.
ticket_create
ChatGPTOpens a support ticket for the IN-STORE team, exactly like the cabinet Support form. For an admin with a selected client it is filed as a note about that client. Without confirm it returns a preview; with confirm: true it makes the change.
ticket_get
ChatGPTOne support ticket with its timeline (id from tickets_list). An admin also gets its change history.
tickets_list
ChatGPTLists the newest support tickets raised by the account (for an admin: by the selected client), with status.
timezone_update
ChatGPTSets the time zone of the signed-in person (profile settings), which the cabinet uses to read and show schedule times. timezone: an IANA name such as America/Toronto, or an empty string to clear it. Without confirm it returns a preview; with confirm: true it makes the change.
video_manage
ChatGPTThe video buttons of a post. action: start (make a video for the post: lengthSec, focus, closingLine; or own script1, script2, script3 all three), remove (scope post or everywhere), publish_youtube (send the finished video to the account's YouTube), youtube_fields (title, description, privacy private|unlisted|public, tags), library_rename (itemId, name), library_delete (itemId, force). Read published_get video first. Without confirm it returns a preview; with confirm: true it makes the change.
voice_calls_get
ChatGPTThe Voice Agent tab: the calls the account's phone agent took. what: status (is the tab ready), calls (the list: q searches the words and summaries, periodDays, success success|failure, hasLead, minDur seconds, sort new|old|long, limit 5 to 60, offset), call (one call with summary, the lead it collected and the transcript: id; the recording stays in the cabinet, whose address the answer gives), stats (totals for periodDays), pulse (how many calls and the newest one). Reading a call here changes nothing: it stays marked as new in the cabinet.
whoami
ChatGPTReturns the STOREKEEPER account this connection acts for: name, role (client or admin), brand, the permissions granted to this connection, and, for an admin, the client currently selected. Call it first when you are unsure whose data you are working with.