Integration

WizCut integration

WizCut edits multicam video podcasts automatically: create a job with the camera angles and audio files, upload them (presigned URLs) or let WizCut import them from a link, and WizCut syncs them, detects who speaks when, works out who is on which camera, cuts between cameras and renders the final MP4. An agent can follow a job through get_job_summary, confirm the speaker mapping, regenerate or replace the cuts, adjust the audio mix and loudness, render, get the video link, approve, retry a failed job, manage the public share link and delete jobs; a person can also review the edit through the review link WizCut returns. Authentication is an API key (starts with wc_live_) sent as a Bearer token, created and revoked on the WizCut settings page. Usage counts input minutes (the duration of the uploaded footage) against the account's plan; when they run out, create_job answers HTTP 403 QUOTA_EXCEEDED.

Connect WizCut How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in WizCut.Built from the 22 WizCut tools below, on the connection you add once.

What you can do with WizCut

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Each badge shows the default agent permission of an endpoint (allowed: an agent calls it without asking; ask first: you are asked before an agent calls it, when the agent or chat asks permission for sensitive actions; never: no agent may call it), which every user can change endpoint by endpoint in their credential settings.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Job, 8 operations

approve_jobApprove jobPOSTAgents: ask firstRecords that a person approved the rendered output. Only works once the job is "complete". Moves it to "approved" and sends the "approved" webhook. Optional: nothing else depends on it.
create_jobCreate jobPOSTAgents: ask firstCreates a multicam podcast editing job from up to 12 camera and audio sources; returns jobId and, per source ID, where to upload the file. Upload them (or use add_source_from_url), then call start_processing.
delete_jobDelete jobDELETEAgents: ask firstPermanently deletes the job and every file it stored: uploads, proxies, speaker-detection audio and the render. Works in any status. Does not give back used minutes. Irreversible: confirm with the user first.
get_jobGet jobGETAgents: ask firstReturns the full job: sources, detected speaker turns, cuts, camera-map proposal, sync and render warnings and freshly signed download URLs (render, uploaded files, audio). Can be large: to poll the status, use get_job_summary.
get_job_summaryGet job summaryGETAgents: ask firstReturns a compact job status for polling: status, render progress, nextAction (what to do next, with the review or download link), output and share links and sync or render warnings. Wait at least 30 seconds between calls.
list_jobsList jobsGETAgents: ask firstLists the account's 50 most recent jobs, newest first, with title, status, creation and update times and a thumbnail per camera. Use it to find a job ID; call get_job_summary or get_job for details.
retry_jobRetry jobPOSTAgents: ask firstRestarts a failed job from the stage that failed (syncing, diarizing or rendering). Check get_job_summary first: when nextAction.type is contact_support, retrying will not help.
start_processingStart processingPOSTAgents: ask firstStarts processing a job once every file is uploaded or imported: audio sync, speaker detection, camera mapping and cuts. Returns at once (status syncing); the work takes minutes: poll get_job_summary or wait for the webhook.

Audio mix, 2 operations

get_audio_mixGet audio mixGETAgents: ask firstReturns a job's audio mix as the next render will play it: each track's level, mute and mic owner, the Balance voices and mic-ducking settings, the detected speakers and the loudness choice.
set_audio_mixSet audio mixPUTAgents: ask firstChanges the soundtrack the next render makes: Balance voices, mic ducking, per-track levels, mutes and owners, loudness. Only fields you send change, but tracks replaces the whole per-track map. Then call start_render.

Job cuts, 2 operations

recut_jobRecut jobPOSTAgents: ask firstRegenerates a ready or complete job's cuts with another pace (relaxed, balanced or dynamic) or new silence-removal settings, without reprocessing, and returns the new cuts. The video is not re-rendered: call start_render.
update_cutsUpdate cutsPATCHAgents: ask firstReplaces a ready or complete job's whole cut list (which camera is shown when) and/or its silence-removal spans: send the complete list. Call start_render afterwards to render the edit.

Job source, 2 operations

add_source_from_urlAdd source from urlPOSTAgents: ask firstAdds one file to a job from an https link (a direct download, or a Google Drive, Dropbox, OneDrive or SharePoint link shared with anyone); WizCut imports it in the background. Returns sourceId; follow it with get_job_summary.
remove_job_sourceRemove job sourceDELETEAgents: ask firstRemoves one source while the job takes files ("created" or "uploading"): a duplicate, a file from the wrong recording, or one whose URL import failed. A running import stops. Irreversible: to get the file back, upload or import it again.

Multipart upload, 2 operations

abort_multipart_uploadAbort multipart uploadPOSTAgents: ask firstCancels a multipart upload and discards the parts uploaded so far (the file must then be uploaded again). Takes the key and uploadId from the source's entry in create_job's uploads.
complete_multipart_uploadComplete multipart uploadPOSTAgents: ask firstCall once every part of a multipart upload has been PUT. Assembles the file and checks its size against the declared `fileSize`. After all files are uploaded, call start_processing.

Job render, 1 operation

start_renderStart renderPOSTAgents: ask firstRenders the job's current cuts into the final video, from ready (or from complete to render again after editing). Returns at once (status rendering): poll get_job_summary until complete, or wait for the complete webhook.

Job share link, 1 operation

update_job_sharingUpdate job sharingPATCHAgents: ask firstTurns a job's public share page on or off, or replaces its link (the old one stops working for good). While sharing is on, anyone with shareUrl can watch and download the finished video. Send share, resetShareLink or both.

Job sync, 1 operation

check_sync_againCheck sync againPOSTAgents: ask firstReads every file's sound again and lines the files up anew, for a job whose syncWarnings list a file with reason unreadable. Returns at once: poll get_job_summary until syncRecheck.running is false, then render.

Job video, 1 operation

get_job_output_urlGet job output urlGETAgents: ask firstReturns location, a freshly signed URL of the job's rendered MP4 valid for a few hours, without downloading the video. Answers HTTP 409 NOT_RENDERED until the job is complete or approved.

Speaker mapping, 1 operation

submit_speaker_mappingSubmit speaker mappingPOSTAgents: ask firstSubmits which speakers each video source shows, while the job is in mapping. WizCut then generates the cuts, moves the job to ready and, for a job created with review false, starts the render. Returns the new cuts.

Upload part url, 1 operation

request_upload_part_urlsRequest upload part urlsPOSTAgents: ask firstFor a source with a multipart upload: returns presigned PUT URLs (valid two hours) for up to 16 more part numbers. PUT each part's bytes to its URL without an Authorization header, then call complete_multipart_upload.

Use WizCut from Claude Code, Cursor or Codex

Connect WizCut in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its WizCut operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect WizCut

Create your workspace
Sign up free. Nothing to install.
Connect WizCut once
Add your WizCut connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.