Free, local, open
Free and open-source AI tools for video creators
46 projects you can download and run yourself, from Whisper transcription to scene detection. Free to download does not mean free for any use, so check the licence.
Definitions
Open source, source available, free plan
Open sourceOSI-approved licence
MIT, Apache-2.0, BSD, GPL, AGPL and similar. You can use, change and share the code under the licence terms. AGPL and GPL add sharing obligations.
Check termsSource available
The code is public but the licence limits use. Examples in the directory: n8n's Sustainable Use License, Remotion's company licence, and non-commercial licences on some agent skills.
Not open sourceFree plan
A paid product with a free tier, such as many cloud editors. See AI tools with a free plan.
Where it runsLocal or self-hosted
Local tools run on your computer. Self-hosted tools run on a server you manage. Either way, check whether the tool calls a paid cloud model or API.
Transcription and subtitles
| Project | What it does | Licence | Runs | Last commit | Watch out |
|---|---|---|---|---|---|
| Auto Subs (tmoroney)tmoroney | Generate speech subtitles inside Resolve, Premiere or After Effects workflows | MITOpen source | On your computer | Oct 2026 | Verify host editor/version compatibility before recommending |
| faster-whisperSYSTRAN | Faster Whisper-based transcription with CTranslate2 | MITOpen source | On your computer | Oct 2026 | Benchmark actual hardware, RAM, word timing and language |
| whisper.cppggml | Run efficient Whisper inference on local hardware | MITOpen source | On your computer | Oct 2026 | Hardware acceleration and language accuracy vary by device |
| WhisperXm-bain | Generate aligned word-level subtitles and diarisation workflows | BSD-2-ClauseOpen source | On your computer | Oct 2026 | Speaker diarisation may need auth and separate model terms |
| Buzz transcriptionchidiwilliams | Desktop transcription / speech-to-text workflow | MITOpen source | Desktop app, On your computer | Nov 2026 | Check privacy and language support with local test |
| Whisper (OpenAI)OpenAI | Speech recognition and subtitles for local post-production | MITOpen source | On your computer | Aug 2026 | Check licence and model weights; speech errors require review |
| pyannote.audiopyannote | Identify speakers and diarise multi-voice vlogs | MITOpen source | On your computer, Cloud | Oct 2026 | Check gated weights, model licenses and speaking overlap |
| Whisper WebUIjhj0517 | Run transcription through local or hosted web UI | Apache-2.0Open source | Self-hosted server, Cloud | Dec 2025 | Self-hosting and exposed web UI security |
| stable-tsjianfch | Improve Whisper word and segment timestamp alignment | MITOpen source | On your computer | May 2026Archived | Inspect drift near music and overlapping speakers |
Editing, scene detection and media processing
| Project | What it does | Licence | Runs | Last commit | Watch out |
|---|---|---|---|---|---|
| FFmpeg / ffprobeFFmpeg | Cut, transcode, crop, concatenate, subtitle and inspect video in scripts | CustomCheck terms | On your computer | Nov 2026 | Not AI itself; invaluable media substrate; encoder licensing depends on build |
| PySceneDetectPySceneDetect | Detect hard cuts and scene boundaries in footage | BSD-3-ClauseOpen source | On your computer | Nov 2026 | Scene cuts are algorithmic, not AI narrative selections |
| VEED OpenEditVEED | Use agent skills to programmatically edit video timeline and render | Apache-2.0Open source | On your computer, Cloud | Nov 2026 | Check render binary license and OS support; published docs differ |
| claude-youtube-editorhassancs91 | Run a talking-head edit pipeline then prepare YouTube publishing | MITOpen source | On your computer, Cloud | Aug 2026 | Require approval before YouTube upload; OAuth security |
| AI Video Editor (timkulbaev)timkulbaev | Local video commands surfaced through CLI and MCP | MITOpen source | On your computer | Feb 2026 | Validate output file controls and sandbox execution |
| FireRed OpenStorylineFireRedTeam | Plan edits conversationally and orchestrate video tools | Apache-2.0Open source | On your computer, Cloud | Jul 2026 | External models and media assets may add cost |
| FunClipModelScope | Transcribe and perform speech-selected cuts from spoken footage | MITOpen source | On your computer | Nov 2026 | Check language and model licensing; local resource requirements |
| OpenChatCut0xsline | Multitrack local editing controlled from agent tools and MCP | AGPL-3.0Open source | On your computer | Oct 2026 | AGPL obligations; do not grant agents uncontrolled access to media |
| Silero VADsnakers4 | Detect segments containing speech before cutting silence | MITOpen source | On your computer | Oct 2026 | Speech detection is not semantic editing; maintain pauses |
| StoryToolkitAIusamaabid | Search transcripts and arrange footage for editing software | GPL-3.0Open source | On your computer | Feb 2025 | Check Resolve/Premiere versions and project safety |
| yt-dlpyt-dlp | Retrieve creator-owned or authorised source video for editing pipelines | UnlicenseOpen source | On your computer | Oct 2026 | Only use where rights and platform terms permit; not a scraping loophole |
| cartcutcartesiancs | Give coding agent access to desktop video timeline via MCP | MITOpen source | Desktop app | Oct 2026 | Third-party catalog only; verify MCP permissions and export controls |
| hypithypit-ai | Use word-level timeline workflows for generative edit variations | CustomCheck terms | On your computer, Cloud | Nov 2026 | Custom license; avoid declaring commercially free |
| beat-synced-editZiadAbdelkarim | Synchronise cuts to musical beats using local Python and ffmpeg | MITOpen source | On your computer | Aug 2026 | Music must be licensed; check rhythmic match quality |
| Monet EditorMonet-AI-Editor | Use local editorctl and MCP inside macOS editor workflow | MITOpen source | Desktop app | Jul 2026Archived | Use sandboxed clips; verify backup and project integrity |
Restoration and effects
| Project | What it does | Licence | Runs | Last commit | Watch out |
|---|---|---|---|---|---|
| GFPGANTencentARC | Restore faces in low-quality video frames | CustomCheck terms | On your computer | Jul 2024 | Face restoration changes identities; not factual forensic proof |
| Real-ESRGANxinntao | Upscale images and frames used in creator video | BSD-3-ClauseOpen source | On your computer | Aug 2024 | AI hallucination may invent fine detail; model licence check |
| CodeFormersczhou | Repair portrait or face footage quality | CustomCheck terms | On your computer | Nov 2025 | Nonstandard NTU-S-Lab licence; model hallucinations possible |
| Practical-RIFEhzwer | Interpolate missing frames for smooth motion | MITOpen source | On your computer | Aug 2026 | Motion interpolation creates synthetic intermediate frames |
| Track-Anythinggaomingqi | Track and segment objects through footage for effects | MITOpen source | On your computer | Dec 2025 | Mask drifting and model weight terms need QA |
Generation, motion graphics and translation
| Project | What it does | Licence | Runs | Last commit | Watch out |
|---|---|---|---|---|---|
| RemotionRemotion | Create repeatable data-driven video templates with React | CustomCheck terms | On your computer, Cloud | Nov 2026 | Not blanket free for every commercial use; licence decision required |
| ComfyUIComfy | Node-based pipelines for image and video generation | GPL-3.0Open source | On your computer, Cloud | Nov 2026 | Resource intensive; partner nodes charge separately |
| OpenCreator (KrillinAI)krillinai | Turn spoken content into subtitled or translated video workflows | Apache-2.0Open source | On your computer, Cloud | Nov 2026 | Review third-party model, voice and translation costs |
| HyperFramesHeyGen | Render HTML/CSS motion graphics to video using programmatic pipeline | Apache-2.0Open source | On your computer | Nov 2026 | Apache-2 per catalog; distinguish SDK framework from avatar SaaS |
| reel-videoAsadMoulviDev | Take an idea to storyboard, keyframes and short video with local review board | MITOpen source | On your computer, Cloud | Aug 2026 | Validate AI images for continuity and provenance |
Automation and agent plumbing
| Project | What it does | Licence | Runs | Last commit | Watch out |
|---|---|---|---|---|---|
| ActivepiecesActivepieces | Connect video steps, approval gates, channels and AI providers | CustomCheck terms | Self-hosted server, Cloud | Nov 2026 | Core community MIT; some business features separate |
| n8nn8n | Build video workflows from intake to review to authorised publishing | CustomCheck terms | Self-hosted server, Cloud | Nov 2026 | NOT OSI open source; commercial redistribution restricted |
| DifyDify (LangGenius) | Build AI-assisted metadata and video script orchestration | CustomCheck terms | Self-hosted server, Cloud | Nov 2026 | Custom licence; verify brand and deployment restrictions |
| KestraKestra | Schedule reliable video ingest, captions and publishing jobs | Apache-2.0Open source | Self-hosted server, Cloud | Nov 2026 | Production complexity; infrastructure upkeep |
| LangflowLangflow | Design LLM and media analysis pipelines with visual components | MITOpen source | Self-hosted server, Cloud | Nov 2026 | Not an editor: requires external media processors |
| Node-REDOpenJS Foundation (Node-RED) | Wire low-code event flows between video systems and APIs | Apache-2.0Open source | Self-hosted server | Oct 2026 | Requires own hosting and proper credential controls |
| OllamaOllama | Run local LLMs to plan scripts, labels and media transformations | MITOpen source | On your computer | Nov 2026 | LLM cannot directly edit video without connected tools |
| WindmillWindmill Labs | Run scripts, scheduled thumbnail jobs and review workflows | CustomCheck terms | Self-hosted server, Cloud | Nov 2026 | Licence varies by components; verify commercial use |
| OpenMontagecalesthio | Orchestrate multi-step video creation from script to rendered story | AGPL-3.0Open source | On your computer | Oct 2026 | AGPL reported by catalog; test setup and current licensing |
| Claude Code Video Toolkitdigitalsamba | Automate scripting, voice tracks and code-driven visual scenes | MITOpen source | On your computer, Cloud | Nov 2026 | Complex chain; measure cost and editability |
| FlowiseFlowiseAI | Build visual assistant flows that query and act on media services | CustomCheck terms | Self-hosted server, Cloud | Aug 2026Archived | Archive status and licensing to verify |
| video-clipforgeJohnson-Jia | Script, narrate and render shorts through a repeatable pipeline | Apache-2.0Open source | On your computer, Cloud | Aug 2026 | Check operational cost and daily scheduling review gates |
Model weights can carry a different licence from the code (for example pyannote diarisation models and some restoration models). Check both.