Skip to main content

web

Answer from the live web: search finds sources, fetch pulls a page back as clean text/markdown (not raw HTML). This is the module for research and question-answering; for driving a page interactively use browser, for raw API calls use http.

Exact tools and parameters

catalog.describe_module web returns the real tools and every parameter (including fetch's content-extraction options), from the live registry.

Tools, at a glance: search, fetch. Extraction is a fetch parameter, not a separate tool.

Config worth knowing​

The default search backend is DuckDuckGo (no API key). Module config can set search_backend, search_fallback, a domain allow/blocklist, cache_ttl, fetch_timeout, max_content_length and per-backend api_keys - see catalog.describe_module web and the config schema it returns.

A working app​

Compiles as-is. The brain uses a Digitorn gateway model (the default - no API key of your own needed); see the note below to use your own key or a local model instead.

yaml
app:
  app_id: web_researcher
  name: Web Researcher
runtime:
  entry_agent: main
agents:
  - id: main
    system_prompt: |
      Answer questions from the live web: search for sources, fetch the best
      pages as clean text, then write the answer with links.
    brain:
      provider: openai
      backend: openai_compat
      model: mimo-v2.5-free   # a model the Digitorn gateway serves
      config:
        api_key: placeholder  # ignored in gateway mode
tools:
  modules:
    web:
      config:
        search_backend: duckduckgo
  capabilities:
    grant:
      - module: web
        tools: [search, fetch]
  • http - raw HTTP verbs and requests
  • browser - interactive browser control