feat: socaity.run() one-liner + 0.1.13 readme and banner#2
feat: socaity.run() one-liner + 0.1.13 readme and banner#2luisfilipemed wants to merge 3 commits into
Conversation
socaity.run("vendor/model", **inputs) resolves the service, installs it on
demand, calls its primary endpoint and returns the result. api_key defaults
to SOCAITY_API_KEY. Verified live against flux-schnell and deepseek-v3.
Replace the outdated readme (which described a simple api and import paths that no longer exist) with accurate, current usage led by socaity.run(). Add a hero banner derived from the socaity_frontend home hero. No em-dashes, brand voice and palette.
Replace fabricated text2speech / voice2voice / audio2face / text2image examples with a real multi-modal pipeline built on socaity.run (deepseek-v3, flux-schnell, kokoro-82m). Verified live end to end.
There was a problem hiding this comment.
I never implemented this .run syntax and it looks very stolen from replicate. Have you tested this?
Also something like face2face is not part of the package has nothing to search in this readme.
What about the sentence replicate backed.
The banner is not fitting to the package just use the simple icon.
I don't like this new version I will discard.
|
@luisfilipemed let's take the new icon you made with the small "py" icon on the bottom; and could you do another run of a readme overhaul? And don't forget to link to the real readme |
There was a problem hiding this comment.
This is okay to have a brand header (need to be updated then also somewhen though) I'd make it smaller.
I can't merge this PR because a new syntax was introduced:
cover = socaity.run("black-forest-labs/flux-schnell", prompt="abstract lime on black, minimal")
voice = socaity.run("jaaari/kokoro-82m", text=script, voice="af_bella")
that is not in the socaity-sdk and is not yet planned.
3 participants
What
socaity.run("vendor/model", **inputs): a one line way to run any catalog model. It resolves the service, installs it on demand, calls its primary endpoint and returns the result.api_keydefaults toSOCAITY_API_KEY.text2img/text2speech) and import paths that no longer exist. The new one leads withsocaity.run().Why
The repo README and examples were stale against the refactored product. No engine changes were needed: 0.1.13 already matches the new backend, so this PR is the convenience entrypoint plus accurate docs and visuals.
Verified live (prod)
socaity.run("black-forest-labs/flux-schnell", prompt=...)returns a 1024x1024 ImageFilesocaity.run("deepseek-ai/deepseek-v3", prompt=...)returns the token chunksNo code changes outside
socaity/__init__.py. Brand voice and palette, no em-dashes.