TencentARC/PhotoMaker

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

10088 stars · 810 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

PhotoMaker is a personalized text-to-image generation method that encodes multiple reference face photos into a stacked ID embedding to generate realistic or stylized human photos of a given identity. It works as an adapter on top of diffusion base models without requiring LoRA fine-tuning, producing results within seconds.

Use cases

• generate realistic photos of a person from reference face images
• create stylized portraits or avatars of a specific identity
• customize text-to-image output to preserve a person's identity
• generate consistent character images across scenes without training a LoRA
• swap identity into different poses and styles via text prompts

When to choose

• you need fast identity-preserving image generation without per-subject training
• you want an adapter compatible with existing diffusion base models and LoRAs
• you need both realistic and stylized human photo generation
• you have multiple reference photos of a person and want high ID fidelity

When to avoid

• you need general-purpose image editing unrelated to human identity
• you cannot run GPU inference locally and don't want to use hosted demos
• you need video or audio generation rather than still images
• your use case requires strict licensing compliance given the non-standard license

Facets

library · maturity stable

image-processing machine-learning stable-diffusion artificial-intelligence image-processing deep-learning python cross-platform text-to-image identity-preservation diffusion personalization cvpr-2024 id-embedding comfyui gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="TencentARC/PhotoMaker")