Let AI agents publish your short-form videos
JSON commands agents can trust
Give Claude, Claude Code, Codex, Cursor, Windsurf, OpenClaw, Hermes Agent, or your own AI agent an official MCP server, npm package, and JSON-first toolchain for discovering accounts, validating a video payload, creating a post, and checking status.
Official MCP server, npm package, and public repo
Taisly Agent Kit is no longer just a setup prompt. Agents can connect through the hosted MCP endpoint, install the npm package for CLI or local MCP, and inspect the public GitHub package before using tools.
OAuth-enabled streamable HTTP endpoint for agents that support remote MCP.
https://app.taisly.com/mcpInstall the CLI, SDK, local stdio MCP server, and bundled SKILL.md.
@taisly/agentPublic package source with README, examples, server metadata, and recipes.
github.com/taisly/agentOfficial MCP server listing for io.github.taisly/agent.
io.github.taisly/agentDiscovery profile, MCP score, and server badge for Taisly Agent Kit.
glama.ai/mcp/servers/taisly/agentGet started in minutes
Use the remote MCP endpoint when your agent supports hosted MCP, or ask a shell-capable agent to install the npm package and complete browser setup. After that, the agent can list connected accounts, validate the post, and publish or schedule it.
Keep a human in the loop: agents should confirm the video, caption, accounts, and schedule before publishing.
Publishing video through MCP?Remote MCP works with public video URLs. Local MCP and CLI can use files on your computer.Read the MCP video upload guidenpm install -g @taisly/agent
taisly setup --agent codex
# open the returned loginUrl
taisly checkin --agent codex
taisly platforms:list
taisly posts:validate --json ./campaign.json
taisly posts:create --json ./campaign.json
taisly posts:status --id <historyId>Give this prompt to your AI agent
Use the same flow with any shell-capable AI agent. Choose your agent above and the commands below will be ready to copy.
Help me set up Taisly Agent Kit.
1. Install the CLI: npm i -g @taisly/agent
2. Run taisly setup --agent codex and send me the login URL so I can finish authentication.
3. Wait for me to finish login in the browser.
4. Run taisly checkin --agent codex.
5. Tell me if Taisly is connected and ready. If not, tell me what failed.
If you cannot install or run the Taisly CLI in this environment, stop and tell me to open Taisly Settings instead.{
"video": "./launch.mp4",
"description": "New product update is live",
"platforms": ["platform-id-1", "platform-id-2"],
"scheduled": "2026-06-14T09:00:00+07:00"
}The payload an agent sends
Agents do not need to understand the whole Taisly app. They only need a video path or URL, a caption, destination platform IDs from platforms:list, and an optional scheduled time.
Built for the agent process
Discover accounts
Agents list connected TikTok, Instagram, YouTube, X, and Facebook accounts before taking action.
taisly platforms:list
taisly platforms:schema --platform TikTokValidate payloads
Agents check video files, destination IDs, captions, and schedule values before creating a live post.
taisly posts:validate --json ./campaign.jsonCreate and track
The package wraps Taisly's video posting API, returns a stable historyId, and lets agents check status.
taisly posts:create --json ./campaign.json
taisly posts:status --id <historyId>
Works with the agents people already use
This is the same simple mental model people expect from modern agent integrations: connect the official MCP server or install the package, finish login in the browser, then let the agent use Taisly tools with human approval before publishing.
Remote MCP
Connect agents that support remote MCP to the OAuth endpoint at https://app.taisly.com/mcp.
npm package
Install @taisly/agent when an agent can run shell commands, needs the CLI, or needs local stdio MCP for files on your machine.
GitHub + SKILL.md
Use the public repository, README, examples, and bundled SKILL.md when your agent can read package instructions before acting.
Command reference for agents
These are the primitives agents can safely compose. The app stays simple for humans, while the package gives agents a predictable process.
taisly setupStart browser-based authentication and return a login URL for the user.
taisly checkinFinish setup after browser approval and store a local agent credential.
taisly platforms:listList connected social accounts before choosing destinations.
taisly platforms:schemaShow platform constraints so the agent can prepare valid video posts.
taisly posts:validateCheck video, caption, account IDs, and schedule before live creation.
taisly posts:createCreate an immediate or scheduled video post and return historyId.
taisly posts:statusRead the posting result later with a stable historyId.
taisly reposts:createCreate a repost rule across connected accounts when the user asks.
Features
Auto Repost
Repost your videos automatically to multiple social media platforms
Auto Post
Post your videos automatically to multiple social media platforms
Schedule Posts
Plan and schedule your content to post at optimal times across platforms
REST API for Video Automation
Programmatically post videos, manage platforms, and track performance. Simple authentication, powerful features.
Perfect for AI Agents
Turn your AI Agents into social media manager. Post to TikTok, Instagram, YouTube, X, Facebook, and more - all from one chat message.
# Get your API key from Settings # Then use it in your requests # List connected platforms curl -X GET https://app.taisly.com/api/private/platform/platforms \ -H "Authorization: Bearer YOUR_API_KEY" # Post a video curl -X POST https://app.taisly.com/api/private/post \ -H "Authorization: Bearer YOUR_API_KEY" \ -F "video=@video.mp4" \ -F "platforms=[\"PLATFORM_ID\"]" \ -F "description=My video description"