Update model from decommissioned llama3-70b-8192 to llama-3.3-70b-versatile

Switch to Openverse API for images (free, no API key)

Switch image source to Google Images with browser-like headers

Use Wikipedia API for images, remove image URLs from prompt

Switch image source from Google to DuckDuckGo

Use Google Images for auto image preview instead of LLM

Add auto image preview for high-confidence results

Update server.js

Update server.js

Switch to OpenAI SDK with IPAI_SK endpoint

Test direct model inference on huggingface.co

List available conversational models

Check raw API tasks endpoints

Query available models from HF tasks API

Test multiple providers

Test multiple models on HF router

Direct router test + better error details

Add HF lib diagnostic endpoint and error details

Check HF env vars

DNS deep check for api-inference

Use @huggingface/inference library for proper API routing

Test Basic auth for HF inference endpoint

Test full HF inference endpoint with POST

Add more HF endpoint tests

Add DNS resolution test and alternate HF endpoints

Fix model defaults and HF URL format

Add network diagnostic endpoint

Add API test endpoint for debugging

Add health check endpoint, fetch timeouts, switch to zephyr-7b-beta

Bulletproof error handling for HF API

Fix HF API: use model in URL path, switch to Phi-3, improve error handling

Add custom DuckDuckGo web search tool

Switch from Anthropic to HF Inference API

Initial commit