TikTok MCP for full organic management

TikTok MCP connects your authorized organic TikTok account to compatible AI clients through Markifact's hosted Model Context Protocol server. It can read account, video, and audience performance and publish videos with approval.

This page is for organic TikTok content and account performance. For paid campaigns, ad groups, creatives, budgets, and Spark Ads, use TikTok Ads MCP. For public competitor research, use TikTok Ads Library MCP.

The current tools are tiktok_get_creator_info, tiktok_get_audience, tiktok_get_report, tiktok_list_posts, tiktok_list_report_fields, and tiktok_publish_post. Together they cover organic discovery, reporting, audience analysis, content listing, and video publishing.

No. It supports both reads and writes. Reports, audience queries, creator checks, and post listings run as reads; publishing a video is a write and waits for your approval in Markifact.

Yes. TikTok MCP can publish a video directly to the connected account or upload it as a draft to the creator's TikTok inbox so the creator can finish it in the TikTok app.

For direct posts, it can set a caption up to 2,200 characters, privacy, comment availability, Duet, Stitch, the cover-frame timestamp, and TikTok's promotional-content or paid-partnership disclosures.

The publish tool accepts a direct video URL or a Google Drive share link set to Anyone with the link. Markifact downloads the file and uploads it to TikTok.

Yes. tiktok_get_creator_info returns the creator name, avatar, allowed privacy levels, maximum post duration, and whether comments, Duet, or Stitch are disabled. Publishing validates the post against those current settings.

Account reporting includes video and profile views, unique viewers, likes, comments, shares, new and lost followers, follower growth, engagement and completion rates, average video engagement, and supported profile actions such as bio-link, email, phone, address, app-download, and lead clicks.

Video reporting can include views, reach, likes, comments, shares, favorites, total and average watch time, full-video-watched rate, new followers, profile views, website clicks, traffic sources, retention, and viewer breakdowns, subject to TikTok returning the field for the account.

Yes. The audience tool supports countries, cities, age groups, genders, and follower active times. TikTok requires at least 100 followers before it returns demographic breakdowns.

Yes. tiktok_list_posts returns up to 500 posts, newest first, with caption, links, thumbnail, media type, duration, whether it is an ad, and the available performance metrics.

Yes. TikTok's configurable report tool uses Markifact's standard date-range, comparison, filtering, ordering, and output controls. You can compare a period with the previous period or another custom range.

TikTok fields vary by level and API availability. tiktok_list_report_fields lets the AI inspect the current metrics and dimensions before building a report instead of relying on an outdated hard-coded list.

The AI can prepare the media, caption, privacy, cover, interaction settings, and disclosures, but tiktok_publish_post is a write operation. Markifact pauses it for explicit approval before anything is uploaded or published.

Yes. A session can take a video from Google Drive, analyze TikTok results, write a report to Sheets, or share a summary in Slack. Every external write is approval-gated.

Markifact's hosted MCP works with compatible clients including Claude, ChatGPT, Cursor, VS Code, and other clients that support authenticated remote MCP servers.

Yes. Markifact works as a TikTok connector for Claude. Add Markifact under Claude's Settings and Connectors, enable the TikTok tools, and Claude can analyze and manage organic TikTok content directly in the chat.