Platform Overview

Contract Snapshot: 72 Agent API operations · 71 MCP tools Version: 1.8.0 | Status: Production | Updated: 2026-05-04 Citedy is a backend service for AI agents. Connect once, configure defaults in the web UI, and let any agent — OpenClaw, Claude Code, Cursor, ChatGPT, or any MCP-compatible client — write posts, generate images, publish to social platforms, scan trends, and produce short-form video. The web UI is for settings and scheduling review. The agent does the work.

Who Is This For

Vibe-coders and solo founders who already use AI agents daily
Micro-companies that want content automation without agency overhead
Developers building agent workflows on top of a reliable content + publish API

Not for: agencies, enterprise, or anyone who wants a web chat interface.

Quick Start for Agents

Register: POST /api/agent/register  (or use SKILL.md onboarding)
Infra health: GET /api/agent/health
Operational status: GET /api/agent/status
Configure defaults: GET /api/agent/settings (update via web UI or PUT API)
Create a post: POST /api/agent/post
Publish it: POST /api/agent/publish

Operational response from /api/agent/status:

summary.operational_status: attention
billing.credits_available: 100
social.connected_count: 0                  -> https://www.citedy.com/dashboard/settings?section=account&tab=integrations
knowledge.documents_count: 0               -> https://www.citedy.com/dashboard/settings?section=blog&tab=knowledge
schedule.scheduled_7d: 0                   -> https://www.citedy.com/dashboard/calendar
content.total_items: 0                     -> /post or /write

Command Reference (Quick)

Agent Command	API Call	Cost
`/post <topic>`	POST /api/agent/post	2-4 credits
`/write <topic>`	POST /api/agent/autopilot	4-8 credits
`/shorts <prompt>`	POST /api/agent/shorts	60-185 credits
`/shorts script`	POST /api/agent/shorts/script	1 credit
`/shorts avatar`	POST /api/agent/shorts/avatar	3 credits
`/shorts merge`	POST /api/agent/shorts/merge	5 credits
`/publish now`	POST /api/agent/publish	0 credits
`/schedule list`	GET /api/agent/schedule	0 credits
`/schedule gaps`	GET /api/agent/schedule/gaps	0 credits
`/scan fast`	POST /api/agent/scan (mode=fast)	2 credits
`/scan deep`	POST /api/agent/scan (mode=deep)	4 credits
`/scan ultra`	POST /api/agent/scan (mode=ultra)	6 credits
`/scan ultra+`	POST /api/agent/scan (mode=ultra+)	8 credits
`/gaps generate`	POST /api/agent/gaps/generate	40 credits
`/gaps list`	GET /api/agent/gaps	0 credits
`/products upload`	POST /api/agent/products	1 credit
`/products list`	GET /api/agent/products	0 credits
`/products search`	POST /api/agent/products/search	0 credits
`/image-style`	PUT /api/agent/image-style	0 credits
`/status`	GET /api/agent/status	0 credits
`/settings`	GET /api/agent/settings	0 credits

These commands describe the full Agent API surface. Public MCP v1 intentionally exposes a smaller tool set (see below).

MCP v1 Tool Surface (Public)

MCP v1 is outcome-first. The goal is to expose stable high-value workflows, not every internal pipeline step.

MCP Tool	Backend API Path	Notes
`agent.health`	`GET /api/agent/health`	Core diagnostics
`agent.status`	`GET /api/agent/status`	Actionable readiness snapshot
`settings.get`	`GET /api/agent/settings`	Read settings in-agent
`schedule.gaps`	`GET /api/agent/schedule/gaps`	Planning visibility
`webhooks.deliveries`	`GET /api/agent/webhooks/deliveries`	Delivery debugging
`products.create`	`POST /api/agent/products`	Upload product doc
`products.list`	`GET /api/agent/products`	List product docs
`products.delete`	`DELETE /api/agent/products/[id]`	Remove product doc
`products.search`	`POST /api/agent/products/search`	Search knowledge chunks
`shorts.generate`	`POST /api/agent/shorts`	Returns processing + poll hint
`shorts.get`	`GET /api/agent/shorts/[id]`	Poll final video status/url
`shorts.script`	`POST /api/agent/shorts/script`	Generate script text
`shorts.avatar`	`POST /api/agent/shorts/avatar`	Generate avatar image
`shorts.merge`	`POST /api/agent/shorts/merge`	Merge clips + subtitles
`gaps.generate`	`POST /api/agent/gaps/generate`	40 credits; optional `favorite_id` (uuid) scopes analysis. Owner-checked → 403 cross-tenant before charge.
`gaps.list`	`GET /api/agent/gaps`	Free; optional `?favorite_id=<uuid>&limit=<1-100>`. Hybrid filter keeps legacy NULL gaps via domain match.

Advanced/deferred for MCP public exposure:

settings.update (safe-field policy needed; web-first by default)
schedule.suggest (heuristic helper, not a critical contract)

Recent changes

2026-05-04 (v1.8.0) — gaps.generate and gaps.list accept optional favorite_id (UUID) to scope output to a specific ai_favorites row. Owner-checked before charge; cross-tenant returns 403. gaps.list also adds limit (1-100). Backwards-compatible: callers that don’t pass favorite_id see legacy behaviour. PR #596.

Authentication

All agent endpoints require:

Authorization: Bearer <AGENT_API_KEY>

Obtain your key at citedy.com/dashboard/settings or via POST /api/agent/register. Key rotation: POST /api/agent/rotate-key

Platform Support Matrix

Platform	Text	Images	Video
Twitter / X	Yes	Yes	Yes
LinkedIn	Yes	Yes	Yes
Facebook	Yes	Yes	Yes
Instagram	Yes	Yes	Reels
YouTube	-	-	Shorts
Reddit	Yes	Yes	No (link only)
Threads	Yes	Yes	-

Getting Started

Platform

MCP

Skills

Agent Guide

Platform Overview

Who Is This For

Quick Start for Agents

Command Reference (Quick)

MCP v1 Tool Surface (Public)

Recent changes

Authentication

Platform Support Matrix

​Who Is This For

​Quick Start for Agents

​Command Reference (Quick)

​MCP v1 Tool Surface (Public)

​Recent changes

​Authentication

​Platform Support Matrix

Who Is This For

Quick Start for Agents

Command Reference (Quick)

MCP v1 Tool Surface (Public)

Recent changes

Authentication

Platform Support Matrix