MCP Tools DirectoryAbout this integration
Local MCP server: recordings or public video URLs into transcript, frames, OCR, wall-clock evidence.
- Transport
- stdio
- Authentication
- none
- Initial setup
- none
- Runtime
- unattended
- Evidence
- documented
- Version
- 0.4.1
- Package
- talkthrough-mcp
- Last compatibility test
- Not independently tested
Connect your agent
talkthrough-mcpLocal model and ffmpeg assets may download on first use, and process_url accesses a user-selected public source. No media was processed and no package or model was executed.
Capabilities: Transcribe local audio and video, Extract and search keyframes and OCR, Anchor media evidence to wall-clock time, Optionally diarize and label speakers, Optionally ingest one public media URL
Connected profiles
Additional details
io.github.korovin-aa97/talkthrough-mcp
Source ↗ · Checked 2026-09-170.4.1
Source ↗ · Checked 2026-09-17CC0-1.0; package licenses are separate
Source ↗ · Checked 2026-09-17pypi
Source ↗ · Checked 2026-09-17talkthrough-mcp
Source ↗ · Checked 2026-09-190.4.1
Source ↗ · Checked 2026-09-19The Python console script is documented for uvx execution as a persistent local stdio server, with optional extras for diarization and public-URL ingestion.
Source ↗ · Checked 2026-09-19The local-file workflow has no account or credential prerequisite. Optional URL ingestion is anonymous and explicitly refuses content that requires sign-in.
Source ↗ · Checked 2026-09-19MCP clients spawn the uvx command over stdio. The server processes local file paths; process_url optionally performs one outbound public-media download.
Source ↗ · Checked 2026-09-19Publisher documentation reviewed; package and integration endpoint not executed or independently security-audited.
Source ↗ · Checked 2026-09-19