Kinocut

Kinocut is a local-first MCP server, Python library, and command-line tool for AI-agent video editing. It is aimed at automation and agency teams turning interviews into captioned vertical clips with a review trail. Choose it for local control and production checks; choose another product if you need hosted editing, formal compliance packages, or a hands-off publishing pipeline.

Overview

Kinocut’s workflow takes local interview media through editing, captioning, reframing, and quality review, then can produce a JSON Video Receipt. The receipt records the request, tools run, resulting changes, guardrails, and items that still require human review. That makes the system more accountable for client work than an opaque one-click clip generator, but it does not remove editorial responsibility or guarantee brand-safe publishing without a person involved.

The published version exposes 203 structured MCP tools, including trimming, merging, subtitling, reframing, and audio normalization. Server-side preflight validation rejects risky or invalid parameters before FFmpeg renders, reducing avoidable failed jobs. Quality gates can check VMAF, loudness, black frames, and audio/video sync, and hold output that falls below configured thresholds.

Key features

Local processing and privacy

FFmpeg and Whisper run on the user’s machine, and the maker says media stays there. The core needs no account or API key and has no analytics or phone-home telemetry. This is a strong fit for teams that want tighter control of interview footage or predictable local automation. It also means the team owns the operational setup rather than handing processing to a hosted service.

Agent and automation workflow

Kinocut integrates with Claude Code, Cursor, generic stdio MCP clients, Python automation, and CI. AI highlight detection, automatic reframing, and social-platform presets support a repeatable short-form pipeline. Receipts and quality gates are particularly useful when multiple edits must be checked before delivery, although the system still expects human review where editorial or brand decisions matter.

Dependencies and boundaries

FFmpeg must be installed separately and available on PATH. Optional Hyperframes support requires Node.js 22 or later. Kinocut is not a hosted multi-tenant SaaS editor, does not include a SOC 2 or ISO certificate package, and does not promise fully automatic brand-safe publishing. Community support is through Forgejo or GitHub issues, with no paid SLA claimed.

Pricing

There is no free trial because the core is free, and no paid tier or usage quota is stated. That makes it easy to adopt for experiments, internal tooling, and agency workflows without a subscription commitment. The trade-off is that infrastructure, installation, and support remain the team’s responsibility; organizations needing contractual service levels or a hosted workspace will need a different model.

Platforms

Kinocut supports API, Linux, macOS, and Windows. Installation guidance covers macOS, Ubuntu/Debian, and Windows. Its local architecture is well suited to development machines, servers, and CI environments where FFmpeg can be managed directly.

Who it's for

Kinocut fits AI-agent builders, automation engineers, and agencies that process interview footage locally and want validation, measurable quality gates, and a receipt for client review. It is also a sensible option for teams that prefer open-source distribution and do not want account-based telemetry. It is less suitable for casual creators seeking an all-in-one browser editor, organizations requiring a paid SLA or formal certification package, and publishers expecting entirely automatic approval and release.

Pros and cons

Pros

• Local-by-default processing: FFmpeg and Whisper run on the user’s machine, with no account, API key, analytics, or phone-home telemetry claimed.
• Broad structured toolset: 203 MCP tools cover core edit, caption, reframing, and audio operations for agent-driven workflows.
• Useful safeguards: Preflight validation and gates for VMAF, loudness, black frames, and sync catch technical problems before delivery.
• Reviewable audit trail: JSON Video Receipts document actions and unresolved human-review items.

Cons

• More setup than a hosted editor: FFmpeg is a separate required dependency, and optional Hyperframes support adds Node.js 22 or later.
• Human review remains necessary: Kinocut does not provide fully automatic brand-safe publishing.
• No paid service commitment: Support is community-based and no paid SLA or formal SOC 2/ISO package is claimed.

Alternatives

For a broader shortlist, see AI Video Repurposing Tools.

• CapsAI is the better choice for Android, iOS, or web use and offers a one-time Starter plan at 49.00 INR, while Kinocut is local and automation-focused.
• Headliner suits users wanting web or mobile access, unlimited video clips on its Forever Free plan, and a hosted workflow.
• NovaClip is a close open-source alternative across API, desktop, web, and self-hosted deployments, with no hosted plan or usage quota stated.
• ScreenApp AI Meeting Notes is preferable when meeting transcription and AI notes matter more than local video-editing controls.
• Flowjin fits hosted social clipping with a one-time 60-minute free credit and 720p output, rather than Kinocut’s local pipeline.
• OpusClip is a hosted option with 60 monthly free credits and up to 1080p clips, but its free exports are watermarked and time-limited.
• reap is suited to mobile and web users who want a free hour per month for clipping and captions, with watermarked 720p exports.
• Splicora offers a hosted credit-and-storage model for teams that prefer managed projects to local infrastructure.

Verdict

Kinocut is a strong choice for AI-agent and agency workflows that need local media handling, technical quality gates, and a receipt that supports human approval. Its free core and broad integrations lower the barrier to building repeatable pipelines. Look elsewhere if you want a polished hosted editor, formal service commitments, or publishing that can proceed without human review and local dependency management.