TOKENLAB SKILLS
Ask for the result. The Skill picks the model.
Skills teach your agent how to get a product photo, a voiceover, or a 3D asset from TokenLab. Each one picks a live model, checks the price, and runs the call.
28 Skills in 7 groups · free, MIT license
Catalog
A Skill for each job
Foundation Skills run every call. Outcome Skills add the method for one kind of result.
- Source
Run a call
tokenlab-runRead the model details, estimate the price, upload files, call, and wait for the result.
- Source
Pick a model
tokenlab-model-pickerChoose the model for a result from the live catalog and current prices.
- Source
Write prompts
tokenlab-promptingPrompt patterns per model family for image, video, music, and speech.
- Source
Product photos
tokenlab-product-photographyTurn a packshot or a brief into studio, lifestyle, and hero shots.
text-to-imageimage-to-imageimage-edit
115 live models - Source
Edit images
tokenlab-edit-imageAdd, remove, replace, or recolor objects, and cut out backgrounds.
image-editimage-to-image
83 live models - Source
Restore and upscale
tokenlab-restore-and-upscaleUpscale, sharpen, and restore old or low-resolution photos.
image-edit
72 live models - Source
Text in images
tokenlab-text-in-imagePosters, packaging, and UI mockups with exact, legible text.
text-to-imageimage-edit
112 live models - Source
Consistent characters
tokenlab-character-consistencyKeep one character or product the same across images and clips.
image-to-imagetext-to-imagereference-to-video
107 live models - Source
Photoreal stills
tokenlab-photoreal-stillsRealistic photos of people, places, food, and scenes.
text-to-image
84 live models - Source
Logos and vectors
tokenlab-logos-and-vectorsLogos, icons, and flat marks, with SVG output.
text-to-imageimage-edit
112 live models
- Source
Animate an image
tokenlab-animate-imageTurn a still image into a short video clip.
image-to-video
46 live models - Source
Product ad video
tokenlab-product-ad-videoTurn a product image or a brief into a short ad clip.
image-to-videoreference-to-videotext-to-video
56 live models - Source
Keyframe transition
tokenlab-keyframe-transitionMake a video that moves from a first frame to a last frame.
start-end-to-video
18 live models - Source
Extend a video
tokenlab-extend-videoContinue a generated clip past its last frame.
video-extension
7 live models - Source
Edit a video
tokenlab-edit-videoRestyle or change an existing clip with a prompt.
video-to-video
8 live models - Source
Motion control
tokenlab-motion-controlMove a character image with the motion of a reference video.
motion-control
2 live models - Source
Talking avatar
tokenlab-talking-avatarTurn a portrait and a script or audio into a talking-head video.
image-to-videotts
57 live models - Source
Cinematic video
tokenlab-cinematic-videoText or image to cinematic shots, directed with camera language.
text-to-videoimage-to-video
54 live models
- Source
Voiceover
tokenlab-voiceoverTurn a script into natural speech with a chosen voice.
tts
11 live models - Source
Transcribe and subtitle
tokenlab-transcribe-and-subtitleAudio or video to transcripts, SRT or VTT subtitles, and translations.
stttranslation
12 live models - Source
Music
tokenlab-musicSongs, instrumentals, and sound effects from a brief.
music
6 live models
- Source
3D assets
tokenlab-image-to-3d-assetTurn an image or text into a downloadable 3D model, such as GLB.
image-to-3dtext-to-3d
8 live models - Source
3D worlds
tokenlab-world-generationTurn text, an image, or a video into an explorable 3D world.
world-generation
4 live models
- Source
Decisions
tokenlab-decisionsTyped yes/no, choice, and score answers for agent branches.
systemone
2 live models
- Source
API integration
tokenlab-api-integrationRunnable TokenLab code for chat, native formats, media, and async tasks.
chat
- Source
Native endpoints
tokenlab-native-endpointsUse the Responses, Messages, Gemini, and decision endpoints.
chatsystemone
- Source
Cost routing
tokenlab-cost-routingCost-aware model routes and fallback chains from live prices.
- Source
OpenAI-compatible migration
tokenlab-openai-compatible-migrationMove an OpenAI-compatible app to TokenLab with a minimal change.
chat
How it works
You ask. The matching Skill answers.
Loads on demand
The agent keeps only each Skill's name and description in context, and reads the full steps when a request matches.
Never hardcodes a model
Defaults refresh daily from the live catalog. Before each call, the Skill reads the model details and the price.
Skills combine
One request can use several Skills, for example a product photo and then an ad video from it.
Install
Add the Skills you need
Install the plugin for every Skill, or add one Skill at a time.
/plugin marketplace add hedging8563/tokenlab-skills
/plugin install tokenlab@tokenlab-toolsRun both commands in Claude Code. The plugin installs every Skill.
codex plugin marketplace add hedging8563/tokenlab-skills
codex plugin add tokenlab@tokenlab-toolsThe plugin installs every Skill.
npx skills add https://github.com/hedging8563/tokenlab-skills --skill tokenlab-run -yWorks with any agent that reads Skill folders. Replace the ID with any Skill from the catalog; start with tokenlab-run.
git clone https://github.com/hedging8563/tokenlab-skillsCopy folders from skills/ into your agent's Skills directory. Include tokenlab-run and tokenlab-model-picker.
In Create
Ready-made setups by industry
Each setup opens in Create with its prompt and settings filled in. Change a few fields, then generate images, video, audio or 3D.
207 Skills
MCP and Skills
Tools and method, used together
TokenLab Skills
Gives the agent the method: which model, which prompt, and how to check the result.
This pageQuestions
Do I need an API key?
Only for paid calls. Installing Skills and reading model details and prices need no key. Set TOKENLAB_API_KEY in the agent's environment before you create images, video, or audio.
What do Skills cost?
The Skills are free and MIT-licensed. You pay the normal TokenLab price for the API calls they make, and the Skill checks the price before each call.
Which agents can use them?
Claude Code and Codex through the plugin, and any agent that reads Skill folders, such as Cursor or OpenCode, through npx Skills or a manual copy.
Do I need MCP too?
No. Skills work alone through the HTTP API. MCP adds callable tools, and the Skills use them when both are installed.