Studio and content
The services behind Studio (formats, skills, the library) and the content
records that Campaigns show and publish. Package pulse.v1; the least role of
each method is in its table. See Studio.
Notes
- Our own model only. Everything here generates on the platform’s own
inference plane through the model router. There is no image-generation tier:
GenerateCreativefor a still image answersUnimplemented. - Skills run as one bounded pipeline: OpenBias on the inputs, the prompt,
TOON output, Go checks, one revision only when a check failed, OpenBias on
the draft, then the plan gate. Drafts land on the board with their scores in
content.metadata. - Publishing queue.
statusispending,publishing,posted,withheldorfailed.reason_statusisrecorded,none,missingorunavailable: never read an emptyreasonas “nothing to report” without checking it. - Speech comes from TIDE’s shared
ttsd(TTSD_URL). When it is switched off, a method that needs speech answersUnavailablerather than silence.
ContentService
pulse.v1.ContentService, declared in pulse.proto.
| Method | Request → Response | Least role | Notes |
|---|---|---|---|
GenerateContent | GenerateContentRequest → stream StreamChunk | editor | GenerateContent streams AI-generated content for a specific channel and topic. |
ListContent | ListContentRequest → ListContentResponse | viewer | ListContent returns paginated content pieces. |
UpdateContent | UpdateContentRequest → Content | editor | UpdateContent edits a suggested draft’s title/body/hashtags/CTA before it is approved or rejected. |
SaveGeneratedContent | SaveGeneratedContentRequest → Content | editor | SaveGeneratedContent persists a piece of studio-generated content so it gets a content_id that Approve/Schedule can act on, and returns the stored Content. |
ApproveContent | ApproveContentRequest → Content | editor | ApproveContent marks content as approved for publication. |
RejectContent | RejectContentRequest → Content | editor | RejectContent marks content as rejected with reviewer notes. |
CreateCreative | CreateCreativeRequest → Creative | editor | CreateCreative creates a new creative asset (image prompt, video script, etc.) |
ListCreatives | ListCreativesRequest → ListCreativesResponse | viewer | ListCreatives lists creative assets for a campaign or content piece. |
UpdateCreative | UpdateCreativeRequest → Creative | editor | UpdateCreative updates a creative asset’s metadata or content. |
DeleteCreative | DeleteCreativeRequest → google.protobuf.Empty | editor | DeleteCreative soft-deletes a creative asset. |
GenerateCreative | GenerateCreativeRequest → stream StreamChunk | editor | GenerateCreative generates a real AI image (via local stable-diffusion.cpp or a configured remote image-gen service) from a prompt and persists it as a Creative — unlike CreateCreative, which only registers metadata for an already-uploaded asset. |
AttachContentMedia | AttachContentMediaRequest → AttachContentMediaResponse | editor | AttachContentMedia attaches the caller’s own uploaded photo or video (an AuditAttachmentService upload) to a piece as a creative, sealed in the media vault, so Instagram can publish it as a post, Reel or Story. On an Instagram piece the file is checked against Instagram’s documented rules first (FAILED_PRECONDITION with the reason in plain words). |
DownloadContent | DownloadContentRequest → DownloadContentResponse | viewer | DownloadContent makes the caller’s piece into files the person can post themselves, with or without a connected account: Instagram = the image(s) and caption.txt (caption, hashtags, alt text) in one .zip; a text channel = its text file. Missing assets are made on demand (an Instagram post’s image). Nothing is published. |
UpdateInstagramMentions | UpdateInstagramMentionsRequest → UpdateInstagramMentionsResponse | editor | UpdateInstagramMentions reads, adds or removes the accounts an Instagram piece mentions. A handle comes from what PULSE has on record (partners, leads, competitors when asked) or from the person; each added one is looked up with Instagram business discovery, and only a found one is tagged when the piece is posted. Nothing is published. |
Also uses StreamChunk, Channel, AssistantToolEvent, AssistantProposal, PaginationRequest, ContentStatus, PaginationResponse.
message GenerateContentRequest {
Channel channel = 1;
string topic = 2;
string tone = 3; // "professional", "casual", "bold", "educational"
string target_audience = 4;
string campaign_id = 5; // optional - link to campaign
string diagnostic_id = 6; // optional - use diagnostic context
string additional_context = 7;
string language = 8; // e.g. "pt-BR"
// The strategy-roadmap item this piece is made for (CampaignService
// GetCampaignPlan): the stored row is linked to it, so its status shows on
// the roadmap card. Optional.
string roadmap_item_id = 9;
}
message Content {
string id = 1;
string user_id = 2;
Channel channel = 3;
string title = 4;
string body = 5;
string hashtags = 6;
string call_to_action = 7;
ContentStatus status = 8;
string campaign_id = 9;
string tone = 10;
string language = 11;
string reviewer_notes = 12;
map<string, string> metadata = 13;
google.protobuf.Timestamp created_at = 14;
google.protobuf.Timestamp updated_at = 15;
google.protobuf.Timestamp published_at = 16;
// Diagnostic this content was generated from, when GenerateContentRequest
// set diagnostic_id — empty if generated without diagnostic context.
string diagnostic_id = 17;
// What the row is as stored ("campaign_brief", "social_post", a channel
// name, ...): the channel enum cannot say that a row is not a post.
string format = 18;
// True when the body is only the brief it was planned from (an audit
// suggestion's one line): nothing was written yet, so it cannot be
// approved or scheduled until it is written.
bool not_written = 19;
}
message ListContentRequest {
PaginationRequest pagination = 1;
ContentStatus status_filter = 2;
Channel channel_filter = 3;
string campaign_id = 4;
}
message ListContentResponse {
repeated Content content = 1;
PaginationResponse pagination = 2;
}
message UpdateContentRequest {
string content_id = 1;
string title = 2;
string body = 3;
string hashtags = 4;
string call_to_action = 5;
}
message ApproveContentRequest {
string content_id = 1;
string reviewer_notes = 2;
}
message SaveGeneratedContentRequest {
string title = 1;
string body = 2;
Channel channel = 3; // optional; a sensible default is applied server-side
string campaign_id = 4; // optional
string diagnostic_id = 5; // optional
// The strategy-roadmap item this piece is made for; see
// GenerateContentRequest.roadmap_item_id. Optional.
string roadmap_item_id = 6;
// The Studio format it was made as (a StudioFormat id, "whitepaper");
// kept in the row's metadata as "studio_format" so the library files it
// under what it is. Optional.
string format = 7;
}
message RejectContentRequest {
string content_id = 1;
string reviewer_notes = 2;
string rejection_reason = 3;
}
// Creative is a visual/media asset associated with content.
message Creative {
string id = 1;
string content_id = 2;
string campaign_id = 3;
string type = 4; // "image", "video", "carousel", "story"
string prompt = 5; // AI generation prompt
string url = 6; // asset URL after generation/upload
string alt_text = 7;
string dimensions = 8; // e.g. "1080x1080", "1920x1080"
map<string, string> metadata = 9;
google.protobuf.Timestamp created_at = 10;
google.protobuf.Timestamp updated_at = 11;
int32 duration = 12; // video duration in seconds (if type == "video")
}
message CreateCreativeRequest {
string content_id = 1;
string campaign_id = 2;
string type = 3;
string prompt = 4;
string alt_text = 5;
string dimensions = 6;
}
message ListCreativesRequest {
string content_id = 1;
string campaign_id = 2;
PaginationRequest pagination = 3;
}
message ListCreativesResponse {
repeated Creative creatives = 1;
PaginationResponse pagination = 2;
}
message UpdateCreativeRequest {
string id = 1;
string prompt = 2;
string alt_text = 3;
string url = 4;
string dimensions = 5;
}
message DeleteCreativeRequest {
string id = 1;
}
message GenerateCreativeRequest {
string prompt = 1;
string content_id = 2; // optional
string campaign_id = 3; // optional
int32 width = 4; // defaults to 1024 if unset (image / legacy video)
int32 height = 5; // defaults to 1024 if unset (image / legacy video)
int32 duration = 6; // requested video duration in seconds (0 = image)
// Platform preset for video creatives. One of: "youtube", "youtube_shorts",
// "instagram_reel", "instagram_feed", "instagram_story". Empty = still image
// (or, with duration>0, a single-scene legacy video at width×height). When
// set, the server picks the platform's resolution/fps and generates a
// multi-scene motion video, ignoring width/height.
string format = 7;
// Number of AI-generated scene frames to animate together (clamped 1–100).
// 0 = the platform preset's default.
int32 scene_count = 8;
// Output frame rate (clamped 12–60). 0 = the platform preset's default (30).
int32 fps = 9;
string avatar_id = 10; // optional, ID of the avatar to use for video generation
string voice_id = 11;
string template_id = 12;
string diagnostic_id = 13;
string target_language = 14; // English, French, Spanish, Portuguese, etc.
}
message AttachContentMediaRequest {
string content_id = 1;
string attachment_id = 2; // the caller's finished photo or video upload
// What the piece is on Instagram: "post" | "reel" | "story"; "" keeps
// what the piece already is.
string instagram_kind = 3;
int32 thumb_offset_ms = 4; // optional Reel cover frame, in milliseconds
string alt_text = 5; // optional photo description (alt text)
}
message AttachContentMediaResponse {
Creative creative = 1;
string instagram_kind = 2; // what the piece now publishes as on Instagram
string missing = 3; // what it still needs before it can publish ("" = ready)
}
message DownloadContentRequest {
string content_id = 1;
}
// DownloadFile is one file at a signed, expiring /media/ URL that downloads
// it under name.
message DownloadFile {
string name = 1;
string url = 2;
string mime_type = 3;
}
message DownloadContentResponse {
// bundle is the whole piece in one file (.zip for Instagram, else the text).
DownloadFile bundle = 1;
// files are the same files one by one.
repeated DownloadFile files = 2;
// notes say what could not be made and why (e.g. text too long to draw).
repeated string notes = 3;
}
// InstagramMention is one account an Instagram piece mentions (@handle in
// the caption) and, once Instagram confirmed it, tags (user_tags).
message InstagramMention {
string handle = 1; // the username, without the @
// Where PULSE got it: "partner", "lead", "competitor", "brief" (the
// person wrote it in what they asked for) or "you" (added in the editor).
string source = 2;
string who = 3; // the name PULSE has on record for it ("" = none)
// "verified": Instagram's business discovery found the account, so it is
// tagged when the piece is posted. "not_checked": it could not be looked
// up (reason says why); it stays in the caption and is not tagged.
string state = 4;
string reason = 5;
string name = 6; // the account's name as Instagram returned it
}
message UpdateInstagramMentionsRequest {
string content_id = 1;
string add = 2; // a handle (or instagram.com link) to mention
string remove = 3; // a handle to stop mentioning
bool recheck = 4; // look every mention up on Instagram again
// Also suggest competitors PULSE found (off unless the person asks).
bool include_competitors = 5;
}
message UpdateInstagramMentionsResponse {
repeated InstagramMention mentions = 1;
// Accounts with a known Instagram handle the piece does not mention yet.
repeated InstagramMention suggestions = 2;
}StudioService
Heavy generative content workflows: whitepapers, podcasts, shorts, etc.
pulse.v1.StudioService, declared in pulse.proto.
| Method | Request → Response | Least role | Notes |
|---|---|---|---|
GenerateWhitepaper | GenerateWhitepaperRequest → stream StreamChunk | editor | |
GeneratePodcast | GeneratePodcastRequest → stream StreamChunk | editor | |
GenerateShorts | GenerateShortsRequest → stream StreamChunk | editor | |
GeneratePresentation | GeneratePresentationRequest → stream StreamChunk | editor | |
GenerateInfographic | GenerateInfographicRequest → stream StreamChunk | editor | |
GenerateReport | GenerateReportRequest → stream StreamChunk | editor |
Also uses StreamChunk, ExtraSource, AssistantToolEvent, AssistantProposal, ExtraSourceKind.
message GenerateWhitepaperRequest {
string topic = 1;
string industry = 2;
string length_preference = 3; // "short", "medium", "in-depth"
// The page offered both for months and sent neither (owner, 2026-10-24:
// Studio "had more relevant details" — now they reach the model).
string tone = 4; // "professional", "authoritative", "technical", "conversational"
string language = 5; // "en", "es", "fr", "de", "zh", "pt"
// Who it is for, what it should make them do (the Studio brief), and the
// pages, links and notes the person attached (Browse / Scrape).
ContentBrief brief = 6;
repeated ExtraSource extra_sources = 7;
}
// ContentBrief is what every Studio piece is for: who reads it, what it should
// make them do, and how it sounds. An empty audience is the latest audit's.
message ContentBrief {
string audience = 1;
string goal = 2; // "awareness", "leads", "sales", "education", "retention"
string call_to_action = 3; // the one thing the reader should do next
string tone = 4; // "professional", "conversational", "bold", "technical"
string language = 5; // ISO 639-1, e.g. "en"
}
message GeneratePodcastRequest {
string source_material = 1;
string host_style = 2; // "conversational", "deep_dive", "news"
int32 target_duration_minutes = 3;
ContentBrief brief = 4;
}
message GenerateShortsRequest {
string topic = 1;
string target_platform = 2; // "youtube", "instagram"
string visual_style = 3;
ContentBrief brief = 4;
int32 duration_seconds = 5; // 15, 30, 60, 90
string hook_style = 6; // "question", "bold_claim", "statistic", "story", "demo"
}
message GeneratePresentationRequest {
string objective = 1;
int32 slide_count = 2;
string brand_colors = 3;
ContentBrief brief = 4;
}
message GenerateInfographicRequest {
string data_source = 1;
string layout_style = 2; // "timeline", "comparison", "process"
ContentBrief brief = 3;
}
message GenerateReportRequest {
string timeframe = 1; // "last_30_days", "Q1_2024"
repeated string highlight_metrics = 2;
string export_format = 3; // "pdf", "csv"
ContentBrief brief = 4;
}StudioFormatService
StudioFormatService — the ONE list of what Studio makes (owner, 2026-10-29: “the navigation bar options from the studio are duplicated or incomplete”). Built on the server from the content-skill catalogue plus the generators this deployment really has; the Studio rail, the “Change format” menu and the library’s categories all read it, so they cannot drift.
pulse.v1.StudioFormatService, declared in studio_formats.proto.
| Method | Request → Response | Least role | Notes |
|---|---|---|---|
ListStudioFormats | ListStudioFormatsRequest → ListStudioFormatsResponse | viewer | No model, no write. |
message ListStudioFormatsRequest {}
message StudioFormat {
// Stable id ("instagram_carousel", "whitepaper", …): the Studio card's
// format, and ListSkills' format filter.
string id = 1;
// The rail and library group ("instagram", "youtube", "linkedin",
// "facebook_x", "blog_email", "long_form", "learning", "visual").
string group = 2;
// What makes it: "skill" (only skills), or a non-skill generator: "post",
// "whitepaper", "podcast", "presentation", "infographic", "report",
// "course", "document" (a PDF), "video".
string generator = 3;
// The format id the generator is called with ("mind_map", "youtube"); "".
string make_as = 4;
// The skills that make it, the first is the default.
repeated string skill_ids = 5;
// Older format ids that now mean this entry (roadmap items, old links,
// older library pieces): "slide_deck" → presentation.
repeated string aliases = 6;
// Content channels a generic "post" of this channel means
// ("CHANNEL_INSTAGRAM" → instagram_post).
repeated string channels = 7;
// False when this deployment cannot make it right now; then
// unavailable_code says why (measured on this server, never assumed).
bool available = 8;
// "media_storage_unavailable" | "".
string unavailable_code = 9;
string unavailable_detail = 10;
// The default skill's purpose (English), for a format the app has no
// words of its own for yet.
string purpose = 11;
}
message ListStudioFormatsResponse {
// In the order the rail shows them, grouped.
repeated StudioFormat formats = 1;
// The groups, in order.
repeated string groups = 2;
}SkillService
SkillService — PULSE’s content skills (grpc/agents/skills/CHANNEL/SKILL/SKILL.md). A skill is one bounded run on PULSE’s own inference: the model writes a TOON draft, deterministic Go checks score it, OpenBias screens the request and the draft, and the plan gate judges it. A draft the guard or the gate withholds is never returned or saved. Nothing a skill writes is published or sent by itself: drafts land on the board for a person to approve, and publishing stays behind PULSE_APPLY_ENABLED plus the channel’s own switch.
pulse.v1.SkillService, declared in skills.proto.
| Method | Request → Response | Least role | Notes |
|---|---|---|---|
ListSkills | ListSkillsRequest → ListSkillsResponse | viewer | ListSkills lists the skills, optionally for one channel and/or format (the Studio card’s format). No model, no write. |
RunSkill | RunSkillRequest → RunSkillResponse | editor | RunSkill runs one skill (by id, or the first for channel+format) and, when save_draft is set, saves a kept draft to the board with its scores. |
message SkillInput {
string name = 1;
string description = 2;
bool required = 3;
}
message SkillSummary {
string id = 1;
string channel = 2;
repeated string content_types = 3;
repeated string formats = 4;
string purpose = 5;
repeated SkillInput inputs = 6;
// Scorer ids the draft is checked by (hook-score, caption-lint, …).
repeated string checks = 7;
// The result is a draft for approval only (comments, replies, DMs).
bool draft_only = 8;
// The channel switch that governs publishing an approved draft ("" = none).
string publish_channel = 9;
}
message ListSkillsRequest {
string channel = 1; // "instagram", "linkedin", … ("" = every channel)
string format = 2; // a Studio format ("post", "shorts", …; "" = any)
}
message ListSkillsResponse {
repeated SkillSummary skills = 1;
}
message RunSkillRequest {
// The skill to run. Empty: the first skill for channel + format.
string skill_id = 1;
map<string, string> inputs = 2;
string channel = 3;
string format = 4;
// Save a kept draft to the board (Campaigns › Board) as pending review,
// with its check scores in the content row's metadata.
bool save_draft = 5;
string campaign_id = 6;
string roadmap_item_id = 7;
// Instagram skills may also propose competitors PULSE found as mentions
// (off by default: partners, leads and handles in the brief only).
bool allow_competitor_mentions = 8;
}
message SkillCheck {
string id = 1;
double score = 2; // 0..100
bool pass = 3;
repeated string findings = 4;
// The scorer could not run; never read as a pass.
string error = 5;
}
message RunSkillResponse {
string skill_id = 1;
string title = 2;
string draft_text = 3;
repeated string hashtags = 4;
// The draft as the model's TOON document (after rewrites).
string output_toon = 5;
repeated SkillCheck checks = 6;
// Withheld by OpenBias or the plan gate: no draft is returned or saved.
bool withheld = 7;
string reason = 8;
// The plan gate: concur, dissent, unable, blocked or off.
string verdict = 9;
string verdict_reason = 10;
bool revised = 11;
// Honest gaps: a revision skipped for time, references left out, research
// that found nothing.
repeated string notes = 12;
// The saved board row ("" when not saved).
string content_id = 13;
// The head of the run's sealed step chain (runi/chain).
string trace_head = 14;
// A carousel draft's slides drawn as 1080x1350 PNGs, in slide order (URLs
// signed for the app). Empty with slide_images_notes saying why.
repeated SlideImage slide_images = 15;
// Why there are no slide images, or what to know about them (the neutral
// palette when PULSE has no brand colours).
repeated string slide_images_notes = 16;
// Each hashtag's size as measured from Instagram, in draft order. Empty when
// the skill does not size hashtags.
repeated HashtagSize hashtag_sizes = 17;
}
// One rendered carousel slide.
message SlideImage {
int32 n = 1; // slide number, from 1
string url = 2; // a signed /media/ URL (expires; never stored)
string alt_text = 3;
int32 width = 4;
int32 height = 5;
}
// One hashtag's size: posts a day, measured from Instagram's hashtag search
// and recent media for the account's own connected professional account.
message HashtagSize {
string tag = 1; // "#sourdough"
// measured | not_measured | unreadable. not_measured: no connected account,
// or the 30-hashtags-a-week budget is spent; unreadable: Instagram refused
// (reason says what to reconnect with).
string state = 2;
// Set only when measured; absent is not 0.
optional double posts_per_day = 3;
string band = 4; // niche | medium | broad (measured only)
string reason = 5; // how it was measured, or why not
int32 sample = 6; // posts the figure counted
}AvatarService
Creating and managing personal avatars from uploaded photos for video gen.
pulse.v1.AvatarService, declared in pulse.proto.
| Method | Request → Response | Least role | Notes |
|---|---|---|---|
CreateAvatar | CreateAvatarRequest → Avatar | editor | |
ListAvatars | ListAvatarsRequest → ListAvatarsResponse | viewer |
message Avatar {
string id = 1;
string name = 2;
string image_url = 3;
google.protobuf.Timestamp created_at = 4;
string voice_url = 5;
}
message CreateAvatarRequest {
string name = 1;
string image_b64 = 2; // base64 encoded image data
string voice_b64 = 3; // base64 encoded audio for voice cloning
}
message ListAvatarsRequest {}
message ListAvatarsResponse {
repeated Avatar avatars = 1;
}CourseService
SERVICE 14: CourseService (HeyGen-style Personalized Courses)
pulse.v1.CourseService, declared in pulse.proto.
| Method | Request → Response | Least role | Notes |
|---|---|---|---|
GenerateCourse | GenerateCourseRequest → Course | editor | |
ListCourses | google.protobuf.Empty → ListCoursesResponse | viewer |
message GenerateCourseRequest {
string topic = 1;
string avatar_id = 2;
int32 modules_count = 3;
string voice_id = 4;
string template_id = 5;
string diagnostic_id = 6;
}
message Course {
string id = 1;
string title = 2;
string description = 3;
repeated CourseModule modules = 4;
}
message CourseModule {
string id = 1;
string title = 2;
string video_url = 3;
string script = 4;
string quiz_json = 5;
}
message ListCoursesResponse {
repeated Course courses = 1;
}PublishingService
SERVICE 16: PublishingService (Cacarejar-style Automation)
pulse.v1.PublishingService, declared in pulse.proto.
| Method | Request → Response | Least role | Notes |
|---|---|---|---|
SchedulePost | SchedulePostRequest → ScheduledPost | editor | |
ListScheduledPosts | google.protobuf.Empty → ListScheduledPostsResponse | viewer | |
ListChannelReadiness | google.protobuf.Empty → ListChannelReadinessResponse | viewer | ListChannelReadiness says, for every channel, whether a post scheduled there now would go out, and if not why: from the same functions the dispatcher gates on, so the answer cannot disagree with what dispatch does. |
RemoveScheduledPost | RemoveScheduledPostRequest → google.protobuf.Empty | editor | RemoveScheduledPost deletes one of the caller’s posts that has not gone out (never a posted or sending one). |
ReschedulePost | ReschedulePostRequest → ScheduledPost | editor | ReschedulePost moves a waiting post to a new time, or queues a missed, withheld or failed one again at that time (its old row stays as history). |
Also uses Channel.
// ChannelReadiness is one channel's publishing answer for the caller.
message ChannelReadiness {
Channel channel = 1;
// publisher is the outward publisher's arming key ("linkedin", "youtube",
// "blog", …), or empty when PULSE has no publisher for this channel: its
// content is drafted and scheduled, never sent.
string publisher = 2;
// publisher_note explains a publisher that exists and cannot post today,
// or the absence of one. Empty when the publisher can post.
string publisher_note = 3;
// armed is the kill switch AND the channel's own switch; armed_reason names
// the layer that decided, in the dispatcher's own words.
bool armed = 4;
string armed_reason = 5;
// owner_gated: the channel publishes with the deployment's credential, so
// only allowlisted owners may publish through it. owner_allowed is whether
// the caller is one; owner_reason says why not.
bool owner_gated = 6;
bool owner_allowed = 7;
string owner_reason = 8;
// The two layers of armed, apart, so a page can say which switch is off
// without parsing armed_reason.
bool kill_switch_on = 9;
bool channel_switch_on = 10;
// account_platform is the integration platform the publisher posts with as
// the caller ("youtube", "instagram", "github"); empty when it posts with
// the deployment's own credential or feed.
string account_platform = 11;
// deployment_ready: the deployment configuration the publisher needs is
// present (configuration only, nothing is called); deployment_reason says
// what is missing.
bool deployment_ready = 12;
string deployment_reason = 13;
// publisher_blocker names why publisher_note is set: "video_generation"
// (the publisher posts video and this deployment cannot make any),
// "no_publisher", or empty.
string publisher_blocker = 14;
// channel_active: the caller keeps this channel Active in Settings (no
// settings saved yet counts as active). Dispatch withholds an inactive
// channel. channel_active_reason is set when the setting could not be read.
bool channel_active = 15;
string channel_active_reason = 16;
// LinkedIn only. credential_source: "integration" (the caller's Settings
// card token), "deployment" (LINKEDIN_ACCESS_TOKEN, used only when the
// caller has no LinkedIn connection) or "" (none). author_kind: "person" or
// "organization"; author_urn may be empty for a person whose id is read
// from LinkedIn at posting time.
string credential_source = 17;
string author_kind = 18;
string author_urn = 19;
}
message ListChannelReadinessResponse {
repeated ChannelReadiness channels = 1;
}
// ScheduledPost is one row of the Publishing queue.
//
// reason/reason_status carry the OUTCOME'S EXPLANATION, following the same
// value+provenance pairing AuditCheck uses for status/notes. A row can stop
// short of publishing for reasons that demand different responses — the kill
// switch was off (withheld), the OAuth token is dead, the platform returned a
// quota error (both failed) — and `status` alone renders all three as one
// indistinguishable "it did not publish".
//
// reason_status exists so an ABSENT reason is never rendered as an empty
// string, which reads like "nothing to report":
//
// "recorded" — reason holds the recorded explanation.
// "none" — the outcome needs no explanation (pending/publishing/
// posted). An empty reason is correct here.
// "missing" — the outcome DEMANDS an explanation (withheld/failed) and
// none was recorded. Say the reason is missing; never blank.
// "unavailable" — the reason could not be read at all (e.g. a pre-migration
// database restored from a replica). The row is still real;
// its explanation is simply not retrievable.
message ScheduledPost {
string id = 1;
string content_id = 2;
Channel channel = 3;
google.protobuf.Timestamp scheduled_at = 4;
string status = 5;
// reason is ONLY ever the recorded explanation, never prose about its own
// absence — that is reason_status's job, so the two never have to be
// disambiguated by inspecting the string.
string reason = 6;
// reason_status is "recorded" | "none" | "missing" | "unavailable".
string reason_status = 7;
// Where it went out ("" = not posted, or the platform gave no link).
string published_url = 8;
// Where the post came from: "campaign" (source_name is its name),
// "checkup" (source_at is when that check-up ran), "autopilot", or ""
// (scheduled by hand, or not known).
string source = 9;
string source_name = 10;
google.protobuf.Timestamp source_at = 11;
// not_connected: the send failed because the channel's account is not
// connected, so PULSE will not post it (the person can download it).
bool not_connected = 12;
// When the post was queued (its row was written).
google.protobuf.Timestamp created_at = 13;
// When it went out, was withheld, failed or was missed; unset while it
// waits, and for rows that ended before this was recorded.
google.protobuf.Timestamp finished_at = 14;
}
message SchedulePostRequest {
string content_id = 1;
Channel channel = 2;
google.protobuf.Timestamp scheduled_at = 3;
}
message ListScheduledPostsResponse {
repeated ScheduledPost posts = 1;
}
message RemoveScheduledPostRequest {
string id = 1;
}
message ReschedulePostRequest {
string id = 1;
google.protobuf.Timestamp scheduled_at = 2;
}BrandPaletteService
BrandPaletteService — the colours and logo PULSE draws an account’s carousels, post images and videos with (release 8). Read from the audited site (its CSS, theme colour and logo) when the audit runs; the owner may override any of them. Text colours are checked for WCAG AA contrast; a colour that fails is replaced by the nearest shade that passes and the note says so. With none known the palette is PULSE’s neutral one.
pulse.v1.BrandPaletteService, declared in brand_palette.proto.
| Method | Request → Response | Least role | Notes |
|---|---|---|---|
GetBrandPalette | GetBrandPaletteRequest → BrandPalette | viewer | The palette as drawn. No model, no write. |
SetBrandPalette | SetBrandPaletteRequest → BrandPalette | editor | Sets (or clears) the owner’s choices and returns the palette as drawn. |
message GetBrandPaletteRequest {}
// One colour as drawn.
message BrandColour {
// "background", "text" or "accent".
string role = 1;
// "#RRGGBB", after the contrast check.
string hex = 2;
// Where it came from: "css", "theme_color", "logo", "owner",
// "legibility" or "neutral".
string source = 3;
// The colour before the contrast check moved it ("" = kept as it was).
string asked_hex = 4;
// Why it was moved, in plain words ("" = kept).
string note = 5;
}
message BrandPalette {
// Background, text and accent, in that order. Empty when neutral.
repeated BrandColour colours = 1;
// The logo at a signed URL ("" = none).
string logo_url = 2;
// "site" or "owner" ("" = none).
string logo_source = 3;
// True when no brand colour is known: PULSE's neutral palette is used.
bool neutral = 4;
// The site the colours were read from ("" = none read yet).
string site = 5;
// What the person should know (the neutral palette, moved colours).
repeated string notes = 6;
// The owner's own choices, as set ("" = from the site).
string own_background = 7;
string own_text = 8;
string own_accent = 9;
// The owner chose no logo: none is drawn, not even the site's.
bool no_logo = 10;
}
message SetBrandPaletteRequest {
// "#RRGGBB"; "" = use the site's (or neutral) colour for that role.
string background = 1;
string text = 2;
string accent = 3;
// A new logo (PNG, JPEG or WebP, at most 2 MB); empty = keep the current.
bytes logo = 4;
// Drop the owner's logo, or their "no logo" (the site's, if any, is used
// again).
bool remove_logo = 5;
// Drop every choice of the owner's.
bool reset_all = 6;
// Draw no logo at all, not even the site's (remove_logo undoes it).
bool no_logo = 7;
}