Phosor
Sign In
Agent skill

Give your agent the whole Phosor pipeline

One folder in your skills directory and your agent can generate video, images, speech and e-commerce product photography through the Phosor API — bring your own LoRA, and manage every job it starts. No integration code, no SDK, nothing to pip install. It follows the agentskills.io open standard, so it behaves the same in Claude Code, OpenClaw and everything else that reads it.

Install

The package is identical everywhere — only the destination directory differs. Pick your agent.

Use ~/.claude/skills/ to make the skill available in every project, or .claude/skills/ inside a repo to scope it to that project.

shell
# 1. Fetch and unpack into the skills directory
mkdir -p ~/.claude/skills
curl -fsSL https://phosor.ai/downloads/phosor-ai-skills.zip -o /tmp/phosor-ai-skills.zip
unzip -o /tmp/phosor-ai-skills.zip -d ~/.claude/skills/
 
# 2. Set your API key
export PHOSOR_API_KEY=your-api-key-here
 
# 3. Verify
python3 ~/.claude/skills/phosor-ai-skills/scripts/phosor_client.py check-key

A working key prints {"status": "valid", ...}. Create one under API Keys in your dashboard.

What the agent gains

Every model, parameter and price below is described inside the skill, so the agent picks the right one without being told.

MiniMax H3 Standard & Turbo video

Flagship

The primary Phosor video path: text-to-video, image-to-video, and Reference-to-Video, each with Standard and Turbo model IDs. 480p and 768p, 4–15 seconds, billed per output second.

Explore MiniMax H3 →

Wan 2.2 video

Text-to-video, image-to-video, speech-to-video lip-sync and motion transfer. Standard and turbo variants up to 1080p.

Image generation & editing

FLUX.2-dev, GPT Image 2, Qwen Image and Z-Image Turbo — text-to-image, image-to-image, and multi-image reference editing.

Bring your own LoRA

Upload .safetensors files or import them from URLs, then generate with them. Multiple LoRAs per request.

Image Studio for e-commerce

Product and model photography suites — selling-point layouts, A+ content, white background, scenes, on-model try-on, plus a set of one-off editing tools.

Speech synthesis

Qwen3-TTS text-to-speech with selectable voices, billed per character. Feed the result straight into speech-to-video for a lip-synced take.

Job management

Submit, poll, cancel mid-flight and browse past generations. Cancelling refunds whatever is still queued.

Drama Studio

Web app

Script-to-screen short drama: bible, screenplay and shot list, then voice, first frames, video and the final cut. Available in the web app today — not yet part of this skill, so an agent cannot drive it.

What's in the package

Three files, no build step. Take them individually if you would rather not unzip — keep the folder layout, the skill references these paths.

SKILL.mdThe skill itself — capabilities, workflows, model tables and constraints the agent readsreferences/api.mdEvery endpoint, parameter, limit and price the skill links out toscripts/phosor_client.pyThe CLI the skill drives — Python 3.7+ standard library only, nothing to pip install

Prefer to call the API directly?

Every endpoint the skill uses is documented on its own page, with request and response examples in cURL, Python and JavaScript.