Last verified
Generate highlight clips from long YouTube, Bilibili, or local videos using AI; runs entirely on your own machine with no subscription required.
Paste a Bilibili or YouTube link, let autoclip score the segments, and export a set of short clips in one session instead of scrubbing through the recording manually. The export presets for YouTube Shorts and Bilibili remove the platform-specific export step.
If the recording has no usable dialogue track (music, gameplay without clear speech structure), the subtitle-based scoring will produce weak results.
Course recordings and lecture replays with dense spoken content score well in autoclip's analysis. The topic segmentation groups related explanations together, making it easier to identify clips that stand alone as content without additional context.
If your course relies heavily on diagram walkthroughs or screen demonstrations with minimal narration, clip quality may be inconsistent.
The CLI accepts local files and outputs JSON results, which fits standard batch ETL patterns. The MCP server exposes the full processing pipeline to MCP-compatible clients, allowing AI agents to trigger autoclip analysis programmatically. Docker Compose deployment makes it straightforward to run as a background service.
If you need frame-level visual analysis or facial recognition to drive clip selection, autoclip's subtitle-only approach is not the right fit.
For teams or individuals currently paying monthly subscriptions to OpusClip, Vizard, or Klap, autoclip removes the recurring cost. Running on a modest server or a personal machine handles most typical workloads. The MIT license permits commercial use without additional licensing fees.
If you process a high volume of videos and need managed infrastructure, uptime guarantees, or transcription quality above what local Whisper provides, a managed service remains the more operationally sound choice.
Creating short clips from long videos is time-consuming by hand. Watching an hour-long recording to find the five minutes worth sharing, then cutting each clip, trimming handles, adding title cards, and exporting at the right resolution for each platform can easily take longer than the original recording. For creators who publish weekly, this editing burden compounds fast.
Paid clipping tools solve the time problem but introduce a different one. Your footage gets uploaded to someone else's server, analyzed by their model, and returned as clips you have no control over. Subscription costs add up, model quality is fixed to whatever the service decides, and if the platform changes pricing or shuts down, your workflow goes with it.
Video files are processed entirely on your machine. Only subtitle text is sent to the language model you select; the raw video never leaves your device. This applies to all three deployment modes: desktop, Docker, and CLI.
Accepts Bilibili links, YouTube links, and local MP4/MOV/MKV files. Videos without subtitles can be transcribed locally using Whisper, installed in one step from the settings page. Existing SRT files can be provided at import time.
Supports Qwen (Tongyi Qianwen), OpenAI-compatible APIs, Google Gemini, Siliconflow, and local models via Ollama or LM Studio. API keys stay in your own settings page. You can switch providers per-project or point at a custom OpenAI-compatible base URL.
The language model reads the full transcript, segments it by topic, and assigns a highlight score to each segment. Clips above the scoring threshold are exported automatically. You can adjust the threshold, reorder clips in the collection view, and edit titles before export.
Built-in export presets cover YouTube Shorts, TikTok, Xiaohongshu, and Bilibili. The exporter can burn subtitles and add title cards. Multi-account Bilibili upload is supported from the desktop app.
The desktop app (macOS Apple Silicon, Windows) bundles Python and FFmpeg with no environment setup. The Docker Compose stack provides a web interface for server or NAS use. The CLI handles batch automation; the MCP server exposes the same pipeline to any MCP-compatible client.
OpusClip is a managed cloud service for AI-powered video clipping, while autoclip is a self-hosted open source tool. Both accept YouTube links and analyze content to extract highlights, but the underlying approach differs on data handling and cost structure.
| Feature | autoclip | OpusClip |
|---|---|---|
| License | MIT | Proprietary |
| Video upload | None (local processing) | Required (cloud) |
| Self-hosting | Yes | No |
| Pricing | Free (bring your own API key) | Paid subscription |
| Model choice | Open (Qwen, Gemini, OpenAI-compatible, local) | Fixed |
| Desktop app | Yes (macOS Apple Silicon, Windows) | Web only |
autoclip is the better choice when data privacy is the deciding factor, when you want to control which language model analyzes your content, or when you process enough videos that a monthly subscription is a meaningful cost. OpusClip is worth considering if you need a fully managed workflow with no infrastructure to maintain, or if your team needs built-in collaboration and sharing features.
Klap is a subscription-based clipping service targeting creators who repurpose long-form video into short clips. Like OpusClip, it handles processing in the cloud.
| Feature | autoclip | Klap |
|---|---|---|
| License | MIT | Proprietary |
| Video upload | None (local processing) | Required (cloud) |
| Self-hosting | Yes | No |
| Pricing | Free (bring your own API key) | Paid subscription |
| Export presets | YouTube Shorts, TikTok, Bilibili, Xiaohongshu | YouTube Shorts, TikTok, Instagram |
| CLI/API access | Yes (CLI + MCP server) | No public API |
autoclip's CLI and MCP server give it an automation edge that Klap does not offer. Developers who want to batch-process videos programmatically or integrate clipping into a larger pipeline will find autoclip more capable here. Klap's polished interface and managed infrastructure remain the better fit for non-technical creators who prioritize ease of use over server management.
Self-hosting autoclip with Docker Compose requires Docker and Docker Compose v2.
```bash
git clone https://github.com/zhouxiaoka/autoclip.git
cd autoclip
cp env.example .env
mkdir -p data logs uploads
docker compose up -d --build
```No. Video files are processed locally on your machine. The only data sent externally is subtitle text, which is forwarded to the language model provider you choose in your settings. If you use a local model via Ollama or LM Studio, nothing leaves your device at all.
autoclip itself is free and MIT licensed. The only cost is the API usage from your chosen language model provider, which the README estimates at a few cents to fractions of a dollar per video depending on length and model. Using a local model via Ollama eliminates API costs entirely.
autoclip includes export presets for YouTube Shorts, TikTok, Xiaohongshu, and Bilibili. It can burn subtitles into clips and add title cards. Multi-account Bilibili upload is available from the desktop app.
autoclip analyzes transcript text to score segments, so it works best with speech-heavy content: interviews, podcasts, courses, and live stream recordings with substantive spoken discussion. Purely visual content, music videos, or gameplay footage without substantial dialogue will produce weaker highlight selections.
The desktop app does not support Linux or Intel Macs. Both platforms can run autoclip via Docker Compose, which provides the same web interface as the desktop app, or via the CLI using Python 3.10+ and FFmpeg installed in a virtual environment.
Alternative to
Grok Bot
AI teammates you own: your keys, your model, your machine.
Alternative to
Amazon Kiro
Spec-driven AI development workflow that keeps context across sessions
Alternative to
Pieces for Developers
Persistent memory for your AI across IDEs
Structured coding workflows for Claude Code, Codex, and Pi
Alternative to
ContextPool
Local project memory for Claude Code, generated entirely offline.
Turn Claude Code into a virtual engineering team with 23 skills
Run autoclip on your own server
Self-host autoclip on a Hostinger VPS.
Deploy on HostingerFeatured
Alternative to
Superhuman
AI email assistant that drafts replies and clears your inbox
Alternative toClaude Managed Agents
Open-source agent deployment API. Self-host or use Omnara Cloud.
Alternative toGrok Bot
AI teammates you own: your keys, your model, your machine.
The projects we add and every new guide. Free, unsubscribe any time.
Claude Code to 
OpenCode
Jira to 
Windshift
Notion to 
AppFlowy
Slack to 
Mattermost
Linear to 
GitLab
Wispr Flow to 
Handy
CapCut to 
Concat
Grok Bot to 
rakazo
Superhuman to 
Inbox Zero
Airtable to 
AITable
Muse to 
OpenClaw
GitHub Copilot to 
Cline
Datadog to 
Netdata
Microsoft Teams to 
Rocket.Chat
Pinecone to 
Qdrant
Zapier to 
n8n
Devin AI to 
OpenHands
Firebase to 
Supabase
1Password to 
KeePassXCConfluence to 
Docmost
Contentful to 
Strapi
Adobe Premiere Pro to 
Shotcut
ChatGPT to 
AnythingLLM
LangSmith to 
Langfuse
