MnrGreg/hermes-venice-web
Hermes plugin that uses Venice.AI web_search and web_scrape tools
This plugin registers Venice AI's /augment endpoints as native web_search and web_extract backends within Hermes. It allows agents to use Venice for web searching and scraping through standard Hermes web tool infrastructure.
- Integrates Venice AI /augment/search and /augment/scrape endpoints
- Supports Brave or Google as search providers
- Compatible with any LLM provider
full readme from github
Hermes Venice Web Provider
Native Hermes web backend powered by Venice AI's privacy-first /augment endpoints.
This plugin registers Venice as a standard web_search / web_extract backend — the same architecture used by Firecrawl, Brave, Tavily, and all other built-in providers. No toolset juggling required.
| Capability | Endpoint | Price |
|---|---|---|
web_search |
POST /augment/search |
$0.01/request |
web_extract |
POST /augment/scrape |
$0.01/URL |
Why Venice
- Zero Data Retention — Brave backend never stores or logs queries.
- Anonymized Google fallback — Venice never links identity to requests.
- Provider-independent — uses standalone
/augmentendpoints, works with any LLM provider (OpenAI, Anthropic, local models, etc.). - Clean LLM-friendly output — structured search results and Markdown extraction.
Installation
One-command (recommended)
hermes plugins install mnrGreg/hermes-venice-web --enable
Manual
mkdir -p ~/.hermes/plugins/web/venice
cd ~/.hermes/plugins/web/venice
# Copy plugin files into this folder
hermes plugins reload
API key
Get your Venice API key at https://venice.ai, then:
echo 'VENICE_API_KEY=vk-XXXXXXXXXXXXXXXX' >> ~/.hermes/.env
Configuration
Select Venice via hermes tools or set it directly in ~/.hermes/config.yaml:
web:
backend: venice
Or use per-capability keys to mix providers:
web:
search_backend: venice # Venice for search
extract_backend: firecrawl # Firecrawl for extract
Keep the default web toolset enabled — Venice powers the standard web_search and web_extract tools directly.
Search provider
Venice routes searches through Brave (default, zero data retention) or Google. Set via environment variable:
echo 'VENICE_SEARCH_PROVIDER=brave' >> ~/.hermes/.env # or "google"
How it works
LLM tool call → web_search / web_extract
↓
web_search_registry
↓
VeniceWebSearchProvider
↓
Venice /augment/search or /scrape
Unlike the legacy tool-plugin approach (which registered separate venice_web_search / venice_web_scrape tools and required disabling the default web toolset), this backend provider integrates natively with Hermes's web tools infrastructure.
Venice web integration approaches
| Approach | How it works | When to use |
|---|---|---|
| This plugin (recommended) | Agent decides per-turn via standard web_search/web_extract |
Agentic workflows; works with any LLM provider |
| Model feature suffix | Append flags to model ID (e.g. default:enable_web_search=on) |
Always-on search with Venice inference |
venice_parameters |
Set enable_web_search / enable_web_scraping in request body |
Direct Venice API calls |
API References
- Search: https://docs.venice.ai/api-reference/endpoint/augment/search
- Scrape: https://docs.venice.ai/api-reference/endpoint/augment/scrape
- Pricing: https://docs.venice.ai/overview/pricing#additional-features
- Hermes web provider docs: https://hermes-agent.nousresearch.com/docs/developer-guide/web-search-provider-plugin
Made for the Hermes community by mnrGreg