polis / DEPLOY.md
AK
docs+build: HF Spaces Dockerfile, recruiter README, deploy guide, MIT license
89d6753
|
Raw
History Blame Contribute Delete
2.38 kB

Deploy Polis — GitHub + Hugging Face

Everything is built and tested. Follow these copy-paste steps. You'll need:


1. Push to GitHub

The repo already has a clean commit history. From inside the polis/ folder:

# create the GitHub repo first at https://github.com/new  (name it "polis", empty, no README)
git remote add origin https://github.com/<YOUR_USERNAME>/polis.git
git branch -M main
git push -u origin main

Replace <YOUR_USERNAME>. If prompted for a password, use a Personal Access Token (GitHub → Settings → Developer settings → Tokens).


2. Deploy to Hugging Face Spaces (Docker)

a) Create the Space

  1. Go to https://huggingface.co/new-space
  2. Owner: you · Space name: polis
  3. SDK: choose DockerBlank
  4. Hardware: CPU basic (free) is enough
  5. Create Space.

b) Push the code to the Space

A Space is just a git repo. From inside polis/:

git remote add space https://huggingface.co/spaces/<YOUR_HF_USERNAME>/polis
git push space main

When prompted, username = your HF username, password = a HF access token with write scope (https://huggingface.co/settings/tokens).

The README.md header (sdk: docker, app_port: 7860) tells Spaces how to build and run the Dockerfile automatically. First build takes ~2–3 minutes.

c) (Optional) Turn on the live LLM engine

By default the Space runs in mock mode ($0) and replays the pre-recorded demo. To make agents think with real OpenAI calls:

  1. In your Space → SettingsVariables and secretsNew secret
  2. Name: OPENAI_API_KEY · Value: your key
  3. (optional) add POLIS_BUDGET_USD = 1.00 to cap spend
  4. Restart the Space.

The header dot turns green and reads LIVE ENGINE · OpenAI.


3. Finishing touches

  • Put the Space URL and GitHub URL into the two placeholders near the top of README.md, then commit and push again.
  • In static/app.js, the "Read the architecture" button opens GitHub — swap the placeholder URL for your repo (search for github.com in that file).
  • Pin the repo on your GitHub profile and add the Space link to your resume/LinkedIn.

That's it — one link shows recruiters a deployed, interactive, agentic AI system.