Search with Perigon

perigon

Try it API

Perigon searches dated news articles from about 150,000 publishers, with story clusters and Wikipedia.

Features2

Modes4

Filters3

Playground

Pinned to Perigon.

Results from Perigon

Features

Perigon delivers 2 of the features on the feature list, each from one or more of its modes.

Feature	Returns
News	News articles on a subject, dated, from news sources; a date window keeps them to the last days.
Encyclopedia	Encyclopedia articles on a subject, read from Wikipedia's licensed copy.

Modes

Perigon has 4 search modes. Choose which of them a search may use under Routing on your account page.

Default search
Semantic
Stories
Wikipedia

Filters Perigon applies

Set them under filters. A filter a provider can’t apply returns 400 unsupported_filter.

Filter	Keeps
`domains`	Only these sites, or with `"mode": "exclude"`, none of them.
`recency`	The last so many days, or a `range` of dates.
`language`	One language, by its code.

request.jsonJSON

{
  "query": "which language has the most native speakers",
  "top_k": 5,
  "requested_provider": "perigon",
  "filters": {
    "domains": { "values": ["en.wikipedia.org"] }
  }
}

Pin a query to Perigon

Over the HTTP API, send "requested_provider": "perigon" with the query and your search key.
In an agent connected over MCP, ask it to search with Perigon.

Results come back in the same shape for every provider, as the docs describe. Leave the field out, and Grounder routes each query to the provider most likely to answer it well.

search.shShell

curl https://api.grounder.net/v1/search \
  -H "Authorization: Bearer $GROUNDER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "which language has the most native speakers",
    "top_k": 5,
    "requested_provider": "perigon"
  }'

import os
import httpx

KEY = os.environ["GROUNDER_KEY"]

r = httpx.post(
    "https://api.grounder.net/v1/search",
    headers={"Authorization": f"Bearer {KEY}"},
    json={
        "query": "which language has the most native speakers",
        "top_k": 5,
        "requested_provider": "perigon",
    },
    timeout=30,
).json()

for hit in r["results"]:
    print(hit["rank"], hit["title"], hit["url"])

const res = await fetch("https://api.grounder.net/v1/search", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.GROUNDER_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    query: "which language has the most native speakers",
    top_k: 5,
    requested_provider: "perigon",
  }),
})

const { results } = await res.json()

claude mcp add --transport http grounder \
  https://api.grounder.net/mcp
# then /mcp, grounder, Authenticate