---
title: TwoCentsHustler AI
emoji: 📈
colorFrom: blue
colorTo: indigo
sdk: docker
pinned: false
license: apache-2.0
---

# TwoCentsHustler AI Space

Local inference on **cpu-basic** (free, unlimited).
Runs `gemma-4-E4B-it-Q4_K_M.gguf` (~2.7 GB) via `llama-cpp-python`.

Fallback provider for the TwoCentsHustler financial news platform.

## Endpoint

`POST /api/ai` — `{ "operation": "analyze"|"summarize"|"cluster", "payload": {...} }`

## Environment Variables

| Variable | Default | Description |
|----------|---------|-------------|
| `GGUF_REPO` | `unsloth/gemma-4-E4B-it-GGUF` | HF repo containing the GGUF file |
| `GGUF_FILE` | `gemma-4-E4B-it-Q4_K_M.gguf` | Quantization variant to load |
| `N_THREADS` | `2` | CPU threads for inference |
| `N_CTX` | `4096` | Context window size |
| `HF_TOKEN` | — | Optional: for gated models |

## Hardware

`cpu-basic` — 2 vCPU, 16 GB RAM.
Inference: ~20-40s per call.