Phosor AI

Generate AI videos (text-to-video, image-to-video) with optional custom LoRA styles via the Phosor AI platform.

For detailed API endpoints, parameters, pricing, and limits, see references/api.md.

Setup

1. 1. Sign up or log in at phosor.ai
2. Go to Settings → API Keys to generate your API key
3. Set it as an environment variable:

CODEBLOCK0

Keep your API key secret. Do not commit it to version control or share it publicly. All API calls are authenticated and billed through this key.

The CLI script is at scripts/phosor_client.py. All commands output JSON to stdout.

Local files used by the CLI:

• - Writes phosor-pending.json to track pending job states locally

Quick Start

Text-to-Video

CODEBLOCK1

Image-to-Video

Two-step flow: upload image first, then submit with the returned S3 key.

CODEBLOCK2

With Custom LoRA

Upload your own LoRA model, then use it in video generation.

CODEBLOCK3

CLI Commands

Job Management

• - submit <prompt> — Submit inference job (T2V/I2V). Options: --width, --height, --num-frames, --fps, --steps, --guidance, --image-url, --lora-id, --lora-scale, --seed, --negative-prompt, INLINECODE14
• INLINECODE15 — Get job status
• INLINECODE16 — Get job result (video URL)
• INLINECODE17 — Poll all pending jobs
• INLINECODE18 — List locally tracked pending jobs
• INLINECODE19 — Get job history. Options: INLINECODE20

File Upload

• - upload-image <file> — Upload image for I2V
• INLINECODE22 — Import image from URL. Options: INLINECODE23
• INLINECODE24 — Upload LoRA (two .safetensors). Options: INLINECODE25
• INLINECODE26 — Import LoRA from URLs. Options: INLINECODE27

LoRA Management

• - loras — List LoRA models. Options: --limit, INLINECODE30
• INLINECODE31 — Get LoRA upload status
• INLINECODE32 — Delete a LoRA model

Utilities

• - check-key — Validate API key
• INLINECODE34 — List available models
• INLINECODE35 — Get quota usage/limits

Key Constraints

Resolutions (exact pairs only)

• - 480p landscape — 854 × 480, max 161 frames
• 480p portrait — 480 × 854, max 161 frames
• 720p landscape — 1280 × 720, max 161 frames
• 720p portrait — 720 × 1280, max 161 frames
• 1080p landscape — 1920 × 1080, max 153 frames
• 1080p portrait — 1080 × 1920, max 153 frames

Frame Alignment

Frames must follow 1 + 4*k where k >= 1 (e.g. 5, 9, 13, ... 81, 85, ...). Server auto-aligns down.

Inference Parameters

• - frames_per_second — default: 16, range: 4–30
• INLINECODE39 — default: 4, range: 4
• INLINECODE40 — default: 1.0, range: 1.0–3.5
• INLINECODE41 — default: 1.0, range: 0.0–2.0

Two-Step Upload Rule

Files must be uploaded before use:

1. 1. Image → upload-image / import-image → returns s3_key → use as INLINECODE45
2. LoRA → upload-lora / import-lora → returns lora_id → use as INLINECODE49

Queue Flow

CODEBLOCK4

The poll command checks all locally-tracked pending jobs and removes completed/failed ones.

Pricing

• - 480p — $0.0002 per frame (0.002 credits)
• 720p — $0.0006 per frame (0.006 credits)
• 1080p — $0.0012 per frame (0.012 credits)
• LoRA multiplier — 1.25x

Exchange rate: 10 credits = $1 USD. Credits pre-deducted, auto-refunded on failure.

Version History

1.0.0 (2026-03-21)

• - Initial release
• Text-to-video and image-to-video generation (Wan 2.2 14B)
• Custom LoRA support (upload, import, use in inference)
• 16 CLI commands
• Model registry with GET /api/v1/models endpoint
• Supported models: wan/2.2-14b/text-to-video, INLINECODE53