Vidtreo

Introduction

Connect Claude, Codex, or any MCP client to VIDTREO — search, watch, record, share, and analyze your recordings by asking in plain words.

VIDTREO MCP

The VIDTREO MCP connects your AI assistant to your VIDTREO workspace. Once connected, you work with your videos and audios by asking in the chat: find a recording, read its transcript, play it, record a new one, share it, or analyze it with Intelligence — without opening the dashboard.

It is a remote server built on the Model Context Protocol, the open standard that AI assistants use to connect to other services. It works with Claude (web, desktop, and Claude Code), Codex, and any MCP client that supports remote servers with OAuth.

https://mcp.vidtreo.com/mcp
You ask: "Find last week's demo calls and tell me what the customer pushed back on"
    → The assistant picks VIDTREO tools: search_videos, get_video_transcript, ask_video
    → The MCP server forwards each call to the VIDTREO API with your sign-in
    → VIDTREO checks your role and the permissions you granted
    → The assistant answers — and can show a player or a recorder right in the conversation

What You Can Do

AreaAsk, for exampleWhat happens
Find"Show me the videos recorded in the Sales environment this month"Searches videos and audios by title, environment, status, and date
Read"Summarize what the candidate said about their last job"Reads the transcript (WebVTT, with timestamps) and answers from it
Watch"Play the onboarding video"Shows a VIDTREO player inside the conversation
Record"Record an audio note called Weekly update"Shows a VIDTREO recorder inside the conversation; the recording uploads to your workspace
Share"Make the demo public and invite ana@acme.com"Changes visibility and download permission, emails personal links, revokes them
Organize"Rename it to Q3 kickoff" · "Delete the test recordings"Renames and deletes videos and audios, reprocesses transcripts
Configure"Create an environment for interviews in Full HD with AI transcription"Creates and updates environments and sets your default one
Analyze"Analyze this call with the Sales discovery lens"Starts an Intelligence analysis and reads its result
Ask"Did the customer mention a budget? Cite the moment"Answers questions about a video from its analysis, with citations
Design lenses"Help me create a lens to screen sales candidates"Interviews you, drafts a complete lens, and publishes it where you choose

Every capability is a tool the assistant can call. The Tools reference lists all of them.

Supported Clients

ClientToolsRecorder and player in the conversation
Claude (web)✓✓ Audio and camera record in the conversation; screen recording continues in VIDTREO
Claude Desktop✓✓ Audio records in the conversation; camera and screen recording continue in VIDTREO
Claude Code✓— The assistant gives you a link to the recorder or the recording in the dashboard
Codex and other MCP clients✓When the client displays MCP Apps; otherwise a link to the dashboard

Connect your assistant has the steps for each client.

Key Concepts

ConceptWhat it is
MCP serverThe VIDTREO endpoint your assistant connects to: https://mcp.vidtreo.com/mcp. It holds no data of its own — every request goes to the VIDTREO API on your behalf
ConnectorWhat Claude calls a connected MCP server. You add VIDTREO once and use it in any chat
ToolOne action the assistant can take, such as search_videos or start_analysis. Each tool is marked read-only, write, or destructive
MCP AppAn interactive page the assistant shows inside the conversation. VIDTREO has two: the recorder and the player
ScopeA permission you grant when you sign in, such as videos:read or intelligence:manage. See Sign-in and permissions
EnvironmentA recording space with its own settings (quality, transcription, audio). Recordings made from the assistant go to the environment you choose or to your default environment

Built-in Safeguards

The tools tell the assistant how to behave, so it works with you instead of deciding for you:

  • It asks before it names anything. Recording titles, environment names, and lens names come from you — the assistant asks instead of inventing them.
  • It confirms before destroying. Deleting a video, an audio, or a lens, revoking a link, or changing who can watch a video are marked as destructive, so clients can ask for your approval, and the assistant only calls them when you explicitly asked.
  • It talks about cost first. Starting an analysis and turning on automatic analysis consume credits; the assistant only does it when you asked and tells you about the cost.
  • Your role always applies. The assistant can do exactly what your VIDTREO role allows — never more. A Viewer who connects the MCP is still a Viewer. You also choose which permissions to grant when you connect.
  • Some settings stay in the dashboard. Your own S3 storage and webhook endpoints are never read or changed through the MCP.
  • Media links stay private. Upload and playback links are short-lived and only reach the recorder and player — the assistant never sees them.

Credits

The MCP uses your workspace's prepaid credits in the same way as the dashboard:

ActionCharged as
Recording from the assistantAny other recording in the environment
start_analysisOne Intelligence analysis, per analyzed minute — see Pricing
set_lens_auto_analyze turned onOne analysis for every new recording in that environment, per lens

Searching, reading transcripts and results, sharing, and configuring use no credits.

Built for teams that already live in their assistant. Recruiters can screen interviews from the chat, sales teams can review calls, and support teams can find the moment a customer reported a bug — all from the conversation where they already work.

Support and Privacy

To disconnect an assistant or revoke its access, see Disconnecting.

Next Steps

On this page