OpenMontage
OpenMontageInc/OpenMontage · AGPL-3.0 · Link only · 526 entries
A whole video production system with no engine of its own. A coding agent reads its pipeline manifests and stage skills, calls its 163 Python tools, checks its own work and stops for your approval, from research to a rendered video.
- Link only
You are free to run it. Code you build on it has to be shared under the same license, so we describe it and link to it, and copy none of it.
- Needs
Python 3.10, FFmpeg, Node 22 and a coding agent to drive it. Nearly everything generative needs a key: its example settings file names about 50 keys and settings across image, video, voice and music services.
- Sends off your machine
We found no telemetry. It reads a .env file, starts a local web board and opens your browser, and calls whichever outside services you have given it keys for. That is the point of 125 of its 163 tools.
- We watched it run
With no keys at all, its own registry reported 36 of 163 tools available.
Its zero-key Remotion demo rendered: a title card, a big-number card, a bar chart and a pie chart.
Its slideshow check told eight identical text cards to revise and scored a varied plan strong.
- Not tried here
No pipeline was run end to end. The 125 keyed tools were read, not run.
- License
AGPL-3.0
You are free to run it. Code built on it has to be shared under the same license, so it is here by link only.
Everything in it
Showing 196 of 526Show everything
Tools
- action_timeline_compilerCompiles a scene plan into a timed list of character actions and poses. Runs on the machine. Works with no keys.
- Asset cacheA cache keyed on a tool's inputs so a repeated paid generation call returns the saved file and is not paid for twice.
- atelier_snapshots.pyRenders one review still per scene of a hand-written composition so the board can show them.
- atlas_3dA 3D model from a text description, through Atlas Cloud. Calls an outside service; keys named in its code: ATLASCLOUD_API_KEY…
- atlas_imageImage generation and editing through Atlas Cloud (Seedream and others). Calls an outside service; keys named in its code…
- atlas_media_smoke.pyAn opt-in paid test run of the Atlas Cloud media tools.
- atlas_refresh_imageOpenAI and Google image models reached through Atlas Cloud. Calls an outside service; keys named in its code: ATLASCLOUD_API_KEY. Needs a…
- atlas_videoVideo generation through Atlas Cloud (Seedance and others). Calls an outside service; keys named in its code: ATLASCLOUD_API_KEY. Needs a…
- audio_energyMeasures how loud a track is over time to find the best point to start the music from, so a quiet intro can be skipped. Runs on the…
- audio_enhanceCleans up speech: noise reduction and loudness levelling with FFmpeg. Runs on the machine; needs ffmpeg. Works with no keys.
- audio_mixerMixes narration, music and effects into one track, with ducking under speech and fades. Runs on the machine; needs ffmpeg. Works with no…
- audio_probeReads the length and stream details of an audio or video file with ffprobe. Runs on the machine; needs ffprobe. Works with no keys.
- auto_reframeConverts a video to another shape, such as wide to tall, keeping the face in frame. Runs on the machine; needs ffmpeg. Works with no keys.
- azure_sttSpeech to text through Azure AI Speech, with word timings. Calls an outside service; keys named in its code: AZURE_SPEECH_KEY. Needs a key…
- azure_ttsText to speech through Azure AI Speech neural voices. Calls an outside service; keys named in its code: AZURE_SPEECH_KEY. Needs a key, a…
- BacklotA local web board that watches the projects folder and shows a production as it runs: stages, script, scene plan, generated assets as a…
- backlot_screenshot_stage.pyBuilds fictional projects and captures the board screenshots used in the README, with Playwright.
- backlot_simulate_run.pyFakes a pipeline run on disk so the Backlot board can be watched updating; useful as a demo of the board.
- backlot_visual_eval.pyA screenshot test for the board: captures fixed views and compares them with saved ones.
- backlot_watch_captures.pyCaptures a board screenshot each time a watched project changes.
- bfl_imageFLUX image generation direct from Black Forest Labs. Calls an outside service; keys named in its code: BFL_API_KEY. Needs a key, a program…
- bg_removeCuts the background out of an image with rembg. Routes to whatever provider is installed or keyed. Needs a key, a program or a download…
- bgm_montage_liteCuts supplied clips to the beat of one music track, through the separate BMTS-Lite program (itself AGPL-3.0-only, by another author). Runs…
- blender_worldAssembles and renders a 3D world in Blender from generated and licensed assets. Runs on your own GPU with model weights you download. Needs…
- bria_fal_imageBria image generation reached through fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a program or a…
- bria_imageBria image generation and editing, direct. Calls an outside service; keys named in its code: BRIA_API_TOKEN. Needs a key, a program or a…
- bvh2clip and add-motionTwo Node scripts that turn a 3D BVH motion-capture file into one of the puppet's flat 2D clips (joint tracks per frame, scaled to a fixed…
- cap_recorderScreen recording through the Cap desktop app, with webcam overlay. Runs on the machine. Works with no keys.
- cartesia_ttsText to speech through Cartesia Sonic. Calls an outside service; keys named in its code: CARTESIA_API_KEY. Needs a key, a program or a…
- character_animation_reviewerChecks a character animation: rig integrity, motion and render readiness, as a report. Runs on the machine. Works with no keys.
- character_rig_rendererRenders a rigged character performance in a browser to frames and video. Runs on the machine. Works with no keys.
- character_spec_generatorTurns an approved character idea into a structured character design file. Runs on the machine. Works with no keys.
- Checkpoint storeWrites and reads a saved state after each stage so a run can resume, and refuses to record a gated stage as done without recorded approval.
- Clip corpusA project-local library of downloaded candidate clips with thumbnails and embeddings, queried offline; the heart of the documentary-montage…
- CLIP embedderA thin wrapper around OpenAI's CLIP ViT-B/32 model that turns images and text into vectors for matching footage to a description. Downloads…
- clip_searchSearches a local library of downloaded clips by meaning, using CLIP embeddings, to fill a described slot. Runs on the machine; needs numpy…
- code_snippetRenders a piece of source code as a highlighted image for an overlay. Runs on the machine; needs pygments, PIL. Needs a key, a program or a…
- cogvideo_videoLocal video generation with CogVideo. Runs on your own GPU with model weights you download. Needs a key, a program or a download before it…
- color_gradeApplies a color grade with FFmpeg filters or a LUT; named looks included. Runs on the machine; needs ffmpeg. Works with no keys.
- comfyui_imageImage generation through a ComfyUI server you run. Runs on your own GPU with model weights you download. Needs a key, a program or a…
- comfyui_musicMusic generation through a ComfyUI server you run. Runs on your own GPU with model weights you download. Needs a key, a program or a…
- comfyui_videoVideo generation through a ComfyUI server you run. Runs on your own GPU with model weights you download. Needs a key, a program or a…
- composition_validatorChecks a planned composition before rendering: catches audio and picture lengths that do not match and similar plan errors. Runs on the…
- Configuration modelLoads config.yaml (model provider, budget mode and caps, approval policy, output defaults) with environment overrides. Budget default is…
- corpus_builderFills a project's clip library: searches the stock and archive sources, downloads, makes thumbnails and indexes. Sources it can reach…
- Cost trackerBudget control: estimate before a call, reserve, record the real spend after; three modes (observe, warn, cap).
- Custom playbook generatorWrites a new schema-valid style playbook when none of the shipped ones fits a brief.
- dashscope_asrSpeech to text through Alibaba's DashScope, with word timings. Calls an outside service; keys named in its code: DASHSCOPE_API_KEY. Needs a…
- dashscope_imageImage generation through Alibaba's DashScope (Qwen Image). Calls an outside service; keys named in its code: DASHSCOPE_API_KEY. Needs a…
- dashscope_ttsText to speech through Alibaba's DashScope (Qwen voices), strong in Mandarin. Calls an outside service; keys named in its code…
- Delivery promiseClassifies what a video has promised to be (for example led by motion or led by stills) at proposal time and locks it, so the render cannot…
- diagram_genDraws a diagram from Mermaid text, with Graphviz or a plain drawing fallback. Runs on the machine. Works in part with no keys.
- direct_clip_searchSearches the same stock and archive sources directly and downloads chosen clips, without building a library first. Routes to whatever…
- doubao_ttsText to speech through Volcengine Doubao Speech, Mandarin first. Calls an outside service; keys named in its code: DOUBAO_SPEECH_API_KEY…
- elevenlabs_soundtrackA soundtrack written to a finished video by ElevenLabs, with stems. Calls an outside service; keys named in its code: ELEVENLABS_API_KEY…
- elevenlabs_ttsText to speech through ElevenLabs. Calls an outside service; keys named in its code: ELEVENLABS_API_KEY. Needs a key, a program or a…
- export_bundlePacks a finished render with its metadata into a folder ready to hand to any platform; it uploads nothing. Runs on the machine. Works with…
- eye_enhanceBrightens and sharpens eyes in talking-head footage. Runs on the machine; needs ffmpeg. Needs a key, a program or a download before it…
- face_enhanceLight face clean-up with FFmpeg filters. Runs on the machine; needs ffmpeg. Works with no keys.
- face_restoreRebuilds degraded faces with CodeFormer or GFPGAN. Runs on your own GPU with model weights you download; needs gfpgan, torch. Needs a key…
- face_trackerFinds and follows a face through footage with MediaPipe, for reframing and face effects. Runs on the machine; needs ffmpeg. Needs a key, a…
- fal_3dA 3D model from text or a picture, and object reconstruction, through fal.ai. Calls an outside service; keys named in its code…
- fal_elevenlabs_musicElevenLabs music generation reached through a fal.ai account. Calls an outside service; keys named in its code: FAL_AI_API_KEY, FAL_KEY…
- fal_elevenlabs_ttsElevenLabs speech reached through a fal.ai account. Calls an outside service; keys named in its code: FAL_AI_API_KEY, FAL_KEY. Needs a key…
- fish_audio_ttsText to speech through fish.audio, including cloned voices by reference. Calls an outside service; keys named in its code…
- flux3_fal_imageFLUX 3 image generation through fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a program or a download…
- flux3_replicate_imageFLUX 3 image generation through Replicate. Calls an outside service; keys named in its code: REPLICATE_API_TOKEN. Needs a key, a program or…
- flux3_videoFLUX video generation through fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a program or a download…
- flux_imageFLUX image generation through fal.ai. Calls an outside service; keys named in its code: FAL_AI_API_KEY. Needs a key, a program or a…
- frame_samplerPulls still frames out of a video with FFmpeg, at set times or intervals. Runs on the machine; needs ffmpeg. Works with no keys.
- freesound_musicSearches and downloads tracks from Freesound.org. Calls an outside service; keys named in its code: FREESOUND_API_KEY. Needs a key, a…
- gemini_fal_imageGoogle's image model reached through fal.ai, generation and editing. Calls an outside service; keys named in its code: FAL_KEY. Needs a…
- gemini_omni_falGoogle's Gemini Omni video model reached through fal.ai. Calls an outside service; keys named in its code: FAL_AI_API_KEY, FAL_KEY. Needs a…
- gemini_omni_videoGoogle's Gemini Omni video model: generates a clip and edits it by conversation. Calls an outside service; keys named in its code…
- gemini_replicate_imageGoogle's image model reached through Replicate. Calls an outside service; keys named in its code: REPLICATE_API_TOKEN. Needs a key, a…
- gemini_ttsText to speech through Google's Gemini speech models, including directed dialogue. Calls an outside service; keys named in its code…
- google_imagenGoogle Imagen image generation. Calls an outside service; keys named in its code: GEMINI_API_KEY, GOOGLE_API_KEY. Needs a key, a program or…
- google_musicMusic generation through Google Lyria. Calls an outside service; needs that service's key or sign-in (see the tool's own install note)…
- google_ttsText to speech through Google Cloud, hundreds of voices across many languages. Calls an outside service; keys named in its code…
- green_screen_compositePlaces keyed talking-head footage over a new background. Runs on the machine; needs ffmpeg, numpy, PIL. Works with no keys.
- green_screen_processorKeys out a green screen. Runs on the machine; needs ffmpeg. Works with no keys.
- grok_imagexAI Grok image generation and editing. Calls an outside service; keys named in its code: XAI_API_KEY. Needs a key, a program or a download…
- grok_ttsNarration through xAI's speech API. Calls an outside service; keys named in its code: XAI_API_KEY. Needs a key, a program or a download…
- grok_videoxAI Grok video generation with its own synchronized sound. Calls an outside service; keys named in its code: XAI_API_KEY. Needs a key, a…
- h3_max_videoThe H3 Max video model on fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a program or a download before it…
- heygen_avatarA presenter video from a HeyGen digital twin. Calls an outside service; keys named in its code: HEYGEN_API_KEY. Needs a key, a program or a…
- heygen_catalog_ttsSpeech from HeyGen's catalog of voices. Calls an outside service; keys named in its code: HEYGEN_API_KEY. Needs a key, a program or a…
- heygen_ttsSpeech from HeyGen's own voice model, with voice discovery. Calls an outside service; keys named in its code: HEYGEN_API_KEY. Needs a key…
- heygen_videoVideo generation through HeyGen, which resells several video models under one key. Calls an outside service; keys named in its code…
- heygen_voice_convertChanges the voice of a recording to another voice while keeping its timing, through HeyGen. Calls an outside service; keys named in its…
- higgsfield_videoVideo generation through Higgsfield Cloud. Calls an outside service; keys named in its code: HIGGSFIELD_API_KEY, HIGGSFIELD_API_SECRET…
- hunyuan_cloud_videoTencent Hunyuan video generation through TokenHub. Calls an outside service; keys named in its code: TENCENT_TOKENHUB_API_KEY. Needs a key…
- hunyuan_imageTencent Hunyuan image generation through TokenHub. Calls an outside service; keys named in its code: TENCENT_TOKENHUB_API_KEY. Needs a key…
- hunyuan_videoLocal Hunyuan video generation. Runs on your own GPU with model weights you download. Needs a key, a program or a download before it works.
- hyperframes_composeRenders an HTML, CSS and GSAP composition to video with HyperFrames, fetched through npx. Runs on the machine; needs npx, ffmpeg. Works…
- ideogram_fal_imageIdeogram generation and regional edits through fal.ai; good at readable lettering. Calls an outside service; keys named in its code…
- ideogram_imageIdeogram generation and precise local edits, direct. Calls an outside service; keys named in its code: IDEOGRAM_API_KEY. Needs a key, a…
- image_genAn older all-in-one image tool; the repo marks it deprecated in favor of the selector. Routes to whatever provider is installed or keyed…
- image_selectorPicks the best available image source for the job, generated or stock, and routes the request to it. Routes to whatever provider is…
- inworld_ttsText to speech through Inworld. Calls an outside service; keys named in its code: INWORLD_API_KEY. Needs a key, a program or a download…
- jimeng_videoVolcengine Jimeng video generation. Calls an outside service; keys named in its code: VOLC_ACCESSKEY, VOLC_SECRETKEY. Needs a key, a…
- kling_avatarA talking presenter clip from one image, through Kling's official API. Calls an outside service; keys named in its code: KLING_API_KEY…
- kling_lip_syncRe-syncs the lips of an existing presenter video to new audio, through Kling's official API. Calls an outside service; keys named in its…
- kling_o3_fal_imageKling image generation and reference edits through fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a…
- kling_o3_fal_videoKling video generation and editing through fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a program or a…
- kling_official_animated_explainer_e2e.pyAn end-to-end check of the Kling provider path; a free dry run by default, paid modes on request.
- kling_official_imageImage generation through Kling's official API. Calls an outside service; keys named in its code: KLING_API_KEY. Needs a key, a program or a…
- kling_official_videoVideo generation through Kling's official API. Calls an outside service; keys named in its code: KLING_API_KEY. Needs a key, a program or a…
- kling_ttsText to speech through Kling's official API. Calls an outside service; keys named in its code: KLING_API_KEY. Needs a key, a program or a…
- kling_videoKling video generation through fal.ai. Calls an outside service; keys named in its code: FAL_AI_API_KEY. Needs a key, a program or a…
- lip_syncLocal lip sync with Wav2Lip, for dubbing. Runs on your own GPU with model weights you download; needs torch, ffmpeg. Needs a key, a program…
- local_diffusionLocal Stable Diffusion image generation on your own GPU. Runs on your own GPU with model weights you download. Needs a key, a program or a…
- ltx_api_videoLTX video generation through the direct LTX API. Calls an outside service; keys named in its code: LTX_API_KEY. Needs a key, a program or a…
- ltx_video_localLocal LTX video generation. Runs on your own GPU with model weights you download. Needs a key, a program or a download before it works.
- ltx_video_modalLTX video generation on a Modal endpoint you host. Calls an outside service; needs that service's key or sign-in (see the tool's own…
- mai_imageMicrosoft MAI image generation on a Foundry deployment. Calls an outside service; keys named in its code: MAI_API_KEY, MAI_ENDPOINT…
- mai_ttsText to speech through Microsoft's MAI Voice models over Azure Speech. Calls an outside service; keys named in its code: AZURE_SPEECH_KEY…
- math_animateMathematical animation with Manim. Runs on the machine; needs manim. Needs a key, a program or a download before it works.
- minicut_imageOpenAI image generation bought through the MiniCut command-line program, price first. Needs the separate MiniCut command-line program…
- minicut_musicMusic (Google Lyria) bought through the MiniCut command-line program: it quotes a price first, then runs on a yes. Needs the separate…
- minicut_music_fitMusic written to fit a finished cut, changing at each edit point, through the MiniCut command-line program. Needs the separate MiniCut…
- minicut_sfxOne sound effect from a line of text, through the MiniCut command-line program. Needs the separate MiniCut command-line program installed…
- minicut_ttsElevenLabs speech bought through the MiniCut command-line program, price first. Needs the separate MiniCut command-line program installed…
- minicut_upscaleCloud video upscale (Topaz) bought through the MiniCut command-line program. Needs the separate MiniCut command-line program installed and…
- minimax_fal_videoMiniMax video generation through fal.ai. Calls an outside service; keys named in its code: FAL_AI_API_KEY, FAL_KEY. Needs a key, a program…
- minimax_imageMiniMax image generation, direct. Calls an outside service; keys named in its code: MINIMAX_API_KEY. Needs a key, a program or a download…
- minimax_videoMiniMax video generation, direct. Calls an outside service; keys named in its code: MINIMAX_API_KEY. Needs a key, a program or a download…
- mistral_sttSpeech to text through Mistral's Voxtral, with segment timings and speaker labels. Calls an outside service; keys named in its code…
- mistral_ttsText to speech through Mistral's Voxtral. Calls an outside service; keys named in its code: MISTRAL_API_KEY. Needs a key, a program or a…
- music_genMusic generation through the ElevenLabs Music API. Calls an outside service; keys named in its code: ELEVENLABS_API_KEY. Needs a key, a…
- music_libraryLooks in a local folder of tracks the user has supplied and lists what is there with lengths. Runs on the machine. Needs a key, a program…
- openai_fal_imageOpenAI's image model reached through fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a program or a…
- openai_imageOpenAI image generation, direct. Calls an outside service; keys named in its code: OPENAI_API_KEY. Needs a key, a program or a download…
- openai_ttsText to speech through OpenAI. Calls an outside service; keys named in its code: OPENAI_API_KEY. Needs a key, a program or a download…
- pexels_imageStock photo search and download from Pexels. Calls an outside service; keys named in its code: PEXELS_API_KEY. Needs a key, a program or a…
- pexels_videoStock video search and download from Pexels. Calls an outside service; keys named in its code: PEXELS_API_KEY. Needs a key, a program or a…
- Pipeline manifest loaderLoads and validates the pipeline YAML manifests.
- piper_ttsFree offline text to speech with Piper; plain but usable narration with no key. Runs on the machine; needs piper. Needs a key, a program or…
- pixabay_imageStock image search and download from Pixabay. Calls an outside service; keys named in its code: PIXABAY_API_KEY. Needs a key, a program or…
- pixabay_musicSearches and downloads music from Pixabay Music with no key. Calls Pixabay's site; no key. Works with no keys.
- pixabay_videoStock video search and download from Pixabay. Calls an outside service; keys named in its code: PIXABAY_API_KEY. Needs a key, a program or…
- Platform profilesNamed output profiles (size, shape, codec) for the main platforms: wide, 4K, tall short-form, square, 21:9.
- Playbook loaderLoads, validates and lists style playbooks, with checks on color palettes, type and contrast.
- Playbook to HyperFrames bridgeTranslates a style playbook into style files a HyperFrames composition can use.
- pose_library_builderBuilds a library of named poses and expressions for a rig. Runs on the machine. Works with no keys.
- Provider scoringRanks the available providers for a job on seven weighted things (fit to the task, quality, control, reliability, cost, speed, continuity)…
- qwen_image_21Local Qwen Image generation and editing from weights you have already downloaded. Runs on your own GPU with model weights you download…
- realtime_audioLive two-way speech sessions with ElevenLabs, Microsoft or xAI; kept apart from narration. Calls an outside service; keys named in its…
- recraft_imageRecraft image generation through fal.ai; aimed at logos and brand graphics. Calls an outside service; keys named in its code…
- refresh_kling_o3_fal_contracts.pyRefreshes the saved request formats for the Kling models on fal.ai from published schemas; submits no jobs.
- remotion_caption_burnBurns word-by-word captions onto a video with a Remotion composition. Runs on the machine; needs ffmpeg, ffprobe. Works with no keys.
- render_demo.pyRenders the curated zero-key demos through the Remotion composer; installs the composer's packages on first run.
- runway_videoVideo generation through the Runway API. Calls an outside service; keys named in its code: RUNWAYML_API_SECRET, RUNWAY_API_KEY. Needs a…
- scaffold_atelier_project.pyCreates the empty project layout for a hand-written ('atelier') Remotion composition: the registration boilerplate and folders, with a…
- Scene pacing checkChecks that the steps of a fake terminal scene land in time with the narration cues, and fails loudly when they drift.
- Scene variation checkerChecks a scene plan for sameness before any asset is made: shot sizes, camera movement, lighting, a flagged high point, texture words…
- scene_detectFinds the cut points in a video. Runs on the machine; needs ffmpeg. Works with no keys.
- screen_capture_selectorChooses between the plain FFmpeg screen recorder and Cap. Routes to whatever provider is installed or keyed. Works with no keys.
- screen_recorderScreen recording with FFmpeg alone. Runs on the machine; needs ffmpeg. Works with no keys.
- seedance_arkSeedance video generation direct from Volcengine Ark. Calls an outside service; keys named in its code: ARK_API_KEY. Needs a key, a program…
- seedance_replicateSeedance video generation through Replicate. Calls an outside service; keys named in its code: REPLICATE_API_TOKEN. Needs a key, a program…
- seedance_videoSeedance video generation through fal.ai; the repo's preferred paid video model. Calls an outside service; keys named in its code…
- seedream_imageSeedream image generation through fal.ai. Calls an outside service; keys named in its code: FAL_AI_API_KEY, FAL_KEY. Needs a key, a program…
- Shot prompt builderTurns structured shot language (camera, movement, subject, lighting and more, in layers) into a prompt tuned for a given video model.
- showcase_cardMakes a closing card for a video with FFmpeg. Runs on the machine; needs ffmpeg, ffprobe. Works with no keys.
- silence_cutterRemoves silent stretches from a recording to make jump cuts. Runs on the machine; needs ffmpeg. Works with no keys.
- Slideshow risk scorerScores a scene plan from 0 to 5 on six things that make a video feel like a slide deck: repetition, decorative visuals, motion without…
- sora_videoOpenAI Sora video generation. Calls an outside service; keys named in its code: OPENAI_API_KEY. Needs a key, a program or a download before…
- Source media reviewInspects footage a user supplies (probe, sampled frames, scenes, transcript) and writes one normalized review file, so planning starts from…
- subtitle_genTurns a timed transcript into subtitle files (SRT, VTT, ASS). Runs on the machine. Works with no keys.
- suno_musicFull songs with vocals from Suno, through a third-party API site. Calls an outside service; keys named in its code: SUNO_API_KEY. Needs a…
- svg_rig_builderBuilds an SVG character rig: parts, pivots and layers. Runs on the machine. Works with no keys.
- talking_headA talking-head video from one face photo and an audio file, locally with SadTalker. Runs on your own GPU with model weights you download…
- threejs_asset_catalogBrings licensed 3D model and texture packs into a local catalog for three.js worlds, keeping the license of each. Routes to whatever…
- threejs_worldBuilds an editable three.js world from a structured description (terrain, regions, camera path) for rendering through HyperFrames. Runs on…
- Tool event logAn append-only log of tool events per project, written whenever a tool runs and read by the Backlot board.
- Tool registryFinds every tool by importing the tools package and reports what is available, by capability and provider. In our zero-key trial it found…
- transcriberLocal speech to text with faster-whisper or WhisperX, word timings included; optional speaker labels with a HuggingFace token. Runs on the…
- transcript_fetcherFetches the existing transcript of a YouTube video. Runs on the machine; needs youtube_transcript_api. Needs a key, a program or a download…
- tts_selectorPicks the best available speech provider for the job by score and routes the request to it. Routes to whatever provider is installed or…
- upscaleUpscales images and video locally with Real-ESRGAN. Runs on your own GPU with model weights you download; needs realesrgan, torch, ffmpeg…
- veo_videoGoogle Veo video generation, through fal.ai or Google directly. Calls an outside service; keys named in its code: FAL_AI_API_KEY. Needs a…
- video_analyzerTakes a reference video apart: transcript, scenes, sampled frames and structure, written up as a brief the agent plans from. Runs on the…
- video_composeThe final render: takes the edit-decision file and assets and renders through FFmpeg, the Remotion composer or HyperFrames, then reviews…
- video_downloaderDownloads a video from a web address with yt-dlp. Runs on the machine; needs yt_dlp. Needs a key, a program or a download before it works.
- video_selectorPicks the best available video source for the job, generated or stock, and routes the request to it. Routes to whatever provider is…
- video_stitchJoins several clips into one with checks, transitions and split-screen layouts. Runs on the machine; needs ffmpeg, ffprobe. Works with no…
- video_trimmerCuts a section out of a video. Runs on the machine; needs ffmpeg. Works with no keys.
- video_understandDescribes frames or clips with a local vision-language model. Runs on your own GPU with model weights you download; needs transformers…
- vidu_videoVidu video generation through fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a program or a download…
- visual_qaAutomatic checks on a rendered video: black frames, frozen stretches and similar faults, from sampled frames. Runs on the machine; needs…
- wan_atlas_videoWan video generation through Atlas Cloud. Calls an outside service; keys named in its code: ATLASCLOUD_API_KEY. Needs a key, a program or a…
- wan_disk_cleanup.shA shell script that reports the disk used by local Wan video generation and, only when given a target and --yes, deletes model weights and…
- wan_fal_videoWan video generation through fal.ai. Calls an outside service; keys named in its code: FAL_KEY. Needs a key, a program or a download before…
- wan_replicate_videoWan video generation through Replicate. Calls an outside service; keys named in its code: REPLICATE_API_TOKEN. Needs a key, a program or a…
- wan_videoLocal Wan video generation. Runs on your own GPU with model weights you download. Needs a key, a program or a download before it works.