Hermes Adds Perplexity as a Backend for Its Web Tools

The integration leaves Hermes’s reasoning model unchanged, but its page extracts are relevant snippets—not a guaranteed complete source record.

By 2 min read
Hermes Adds Perplexity as a Backend for Its Web Tools
Hermes Adds Perplexity as a Backend for Its Web Tools

Listen to this story

The audio brief

About 1:30
0:001:30
Read transcript
Hermes can now use Perplexity to search the web and extract page content, without changing the model that handles reasoning and writing. That gives the agent a cleaner division of labor: one provider retrieves information, while Hermes’s existing model interprets it and produces the response. The integration arrives in Hermes version zero point twenty-one point one and later. It uses a separate Perplexity API key and sends requests directly over HTTP, so users do not need Perplexity’s SDK, command-line tool, or MCP server. Web search returns ranked results, with titles, URLs, and short descriptions, and Hermes caps that search at twenty results. The important limitation is in extraction. Web extract returns passages relevant to a requested URL, not necessarily a complete copy of the page. That means an agent can miss a crucial detail. Perplexity’s own example shows an excerpt referring to “the new function” without naming it. The missing name is not evidence for the agent to reconstruct from memory or inference. Configuration also has a trap. Existing search-specific or extraction-specific settings override the shared web backend, so adding a Perplexity key may not redirect either tool. For a real test, Hermes recommends disabling caching and keyless fallback, then checking that a web-search request returns usable results from Perplexity. The constraint worth watching is simple: better retrieval does not remove the need to verify what the extract actually contains.

Story brief

3 key points

Hermes now lets users separate retrieval from reasoning by routing its existing web_search and web_extract tools to Perplexity, while leaving the agent’s model provider unchanged. The integration is available from Hermes v0.21.1, uses direct HTTP rather than Perplexity’s SDK, CLI, or MCP server, and supports a separate API key. Its main limitation is evidentiary: extraction returns relevant passages, not necessarily...

  1. 01

    Perplexity support starts in Hermes v0.21.1; v0.21.0 does not include it.

  2. 02

    web_search is capped at 20 results and uses short context for descriptions.

  3. 03

    web_extract may omit key details, so agents should not fill gaps from memory or inference.

Hermes Agent can now use Perplexity as the backend for its existing web search and page-extraction tools without changing the model it uses for reasoning and writing. The new option creates a clean split between retrieval and the agent model, but it comes with an evidence boundary: extraction returns relevant passages, not a complete copy of a source page.

The provider is included in Hermes v0.21.1 and later; v0.21.0 does not include it. Users can keep their current model provider for reasoning and writing while configuring a separate Perplexity API key for search and extraction.

One agent, separate jobs

The change affects Hermes’s web tools rather than the model running the agent. Hermes sends requests to Perplexity directly over HTTP, so the setup does not require Perplexity’s SDK, command-line interface, or an MCP server.

Relevant text can still be incomplete

Hermes limits Perplexity search requests to 20 results and uses short search context for result descriptions. Extraction works differently: it returns selected material relevant to a URL, which can leave out details found on the original page.

That distinction matters when an agent is asked to make a source-grounded claim. Perplexity’s documentation gives an example in which an extract preserves a reference to “the new function” but omits the function’s name. The missing name is not established by the retrieved excerpt, so it should not be supplied from memory or guesswork.

Old settings can override the new backend

Hermes offers a tools setup wizard as well as manual configuration. For existing installations, search-specific and extraction-specific settings take precedence over the shared web backend, so adding an API key alone may not redirect either tool to Perplexity.

Manual settings can route both tools to Perplexity
web:
  backend: perplexity
  search_backend: perplexity
  extract_backend: perplexity

Perplexity recommends disabling caching and keyless fallback during testing. Otherwise, a cached response or another provider could satisfy a request after a Perplexity error; a meaningful check shows a web_search call and usable returned results.

Sources

  1. docs.perplexity.aiPerplexity web search in Hermes - Perplexity

Loading discussion...