Docs
Models

Service Catalog

Socaity ships typed Python clients for every service in the catalog. The import looks the same for every service — what changes is whether you need to install it first.

This page lists the most-used services grouped by category. The full live catalog with playgrounds, pricing, and runtime status lives at socaity.ai/APIs/overview.

The import pattern

Official services (face2face, speechcraft, audio2face) are auto-installed when you run socaity. Just import them. Unofficial services (community services and Replicate-backed ones like flux_schnell) need a one-time install via socaity --install <name>, then they import the same way.

# Official service — auto-installed, just import
from socaity import speechcraft

# Unofficial service — install once, then import the same way
# socaity --install flux_schnell   (Coming Soon)
from socaity import flux_schnell

# Identical calling pattern for both
flux = flux_schnell()
images = flux(prompt="a cat on mars").get_result()
images[0].save("output.png")

All services authenticate with SOCAITY_API_KEY. The SDK handles authentication for you — official or unofficial, the calling pattern is identical. The socaity --install command is Coming Soon; see the SocaitySDK CLI reference for the current install workflow.

Featured services

A hand-picked selection across vision, speech, image and text. Official services first, then community / Replicate-backed ones. Click any card to open the live service page on socaity.ai.

VISION

official

face2face

by socaity

Face swap and enhancement. Insightface backbone + GFPGAN restoration. Single-shot, no training.

face2face

SPEECH

official

speechcraft

by socaity

TTS + voice cloning. 13 languages, emotion markers, sub-second synthesis. Bark-based.

speechcraft

IMAGE

community

flux-schnell

by black-forest-labs

Fast text-to-image. Four-step diffusion, 1MP output, ~1s per image on a warm worker.

flux_schnellsocaity --install flux_schnell

IMAGE

community

sdxl

by stability-ai

Stable Diffusion XL. The reference quality model — slower than flux-schnell but more controllable.

sdxlsocaity --install sdxl

IMAGE

community

stable-diffusion

by stability-ai

The original Stable Diffusion. Cheapest entry-point for text-to-image when SDXL is overkill.

stable_diffusionsocaity --install stable_diffusion

LLM

community

deepseek-v3

by deepseek-ai

Reasoning-tuned LLM. Use for math, code generation, and step-by-step problem solving.

deepseek_v3socaity --install deepseek_v3

LLM

community

deepseek-v3

by deepseek-ai

General-purpose Chinese-English LLM. Stronger on long-context and structured output than R1.

deepseek_v3socaity --install deepseek_v3

LLM

community

meta-llama-3-70b-instruct

by meta

Llama 3 70B instruction-tuned. The open-weights baseline for production chat workloads.

meta_llama_3_70b_instructsocaity --install meta_llama_3_70b_instruct

LLM

community

gpt-4o

by openai

OpenAI flagship multimodal. Available via Replicate routing — pay per token, no separate OpenAI key needed.

gpt_4osocaity --install gpt_4o

SPEECH

community

whisper

by openai

Speech-to-text. Multi-language transcription with timestamps and language detection.

whispersocaity --install whisper

SPEECH

community

musicgen

by meta

Text-to-music generation. Conditioned on a text prompt, optional melody reference.

musicgensocaity --install musicgen

VISION

community

sam-2

by meta

Segment Anything 2. Click-to-segment for images and video. Useful for masking and object isolation.

sam_2socaity --install sam_2

Browse the full catalog

The featured list above covers the most common entry points. The complete catalog — every official service plus the full community / Replicate index, with playgrounds, current pricing, and runtime status — is on the main platform.

Open API catalog on socaity.ai Python SDK reference

Job System

API Keys