Integration

Perigon integration

Perigon is a real-time news intelligence API. Search and filter enriched news articles, story clusters and their history, media sources, journalists, companies, people, topics and Wikipedia pages, run semantic vector search, get an AI summary of matching articles, and manage AI monitors (events, newsletters, summaries), watchlists, source groups and webhook contact points. Journalist email addresses, webhook URLs, webhook signing secrets and contact point email addresses are withheld from every answer, and the key-status check of the usage route (which takes API keys as input) is not exposed. Authentication is a Perigon API key (created under Developers, API keys) sent as a Bearer token.

Connect Perigon How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Perigon.Built from the 45 Perigon tools below, on the connection you add once.

What you can do with Perigon

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

All, 8 operations

search_articlesSearch articlesGETSearches enriched news articles by keywords (Boolean), source, date, location, people, companies, topics and sentiment; returns articles with entities, sentiment and story clusterId. page starts at 0.
search_companiesSearch companiesGETSearches companies Perigon tracks by name, domain, ticker symbol, industry and more, with Boolean logic and filters such as country, exchange, employee count and IPO date. page starts at 0.
search_journalistsSearch journalistsGETSearches journalists by name, outlet, topics and other attributes (230,000+ worldwide); returns profiles with top topics and sources. Journalist email addresses are withheld. page starts at 0.
search_peopleSearch peopleGETSearches public figures in Perigon's entity database (650,000+, from Wikidata) by name, Wikidata id or occupation; wikidataId feeds personWikidataId of search_articles. page starts at 0.
search_sourcesSearch sourcesGETSearches the 200,000+ media sources by domain, name, country, traffic, media bias and more; returns source profiles. Domains feed the source filter of search_articles. page starts at 0.
search_storiesSearch storiesGETSearches stories (clusters of related articles) with summaries, key points, sentiment and top people, companies, topics and countries. Pass a story id as clusterId to search_articles for its articles.
search_topicsSearch topicsGETSearches the topics Perigon tags articles with, by name, category or subcategory; topic names feed the topic filter of search_articles. page starts at 0.
search_wikipediaSearch wikipediaGETSearches and filters Wikipedia pages Perigon indexes (by text, title, Wikidata id, page views, revision date); returns pages with sections and metadata. page starts at 0.

Contact point, 5 operations

create_contact_pointCreate contact pointPOSTCreates a WEBHOOK contact point that Perigon will POST monitor results to; returns its uuid for contactPointIds of create_monitor or update_monitor. Only WEBHOOK can be created here.
delete_contact_pointDelete contact pointDELETEDeletes a contact point; one with past notifications is archived instead. Perigon advises also removing its uuid from monitors' contactPointIds (update_monitor) to stop delivery.
get_contact_pointGet contact pointGETGets one contact point (uuid from list_contact_points or create_contact_point): name, type, status and verification. Its webhook URL, signing secret and email are withheld.
list_contact_pointsList contact pointsGETLists your organization's webhook and email contact points (where monitors deliver results). sortBy and sortOrder are required. Webhook URLs, signing secrets and emails are withheld.
update_contact_pointUpdate contact pointPATCHChanges a contact point's name, status or webhook URL; only the fields sent change. A new webhook URL receives every later delivery of the monitors that use this contact point.

Monitor, 5 operations

archive_monitorArchive monitorDELETEArchives a monitor (status ARCHIVED): it stops running and is hidden from list_monitors unless status ARCHIVED is asked for. activate_monitor only restarts DRAFT or STOPPED monitors.
create_monitorCreate monitorPOSTCreates an AI news monitor and returns it with its uuid. status is required: DRAFT does not run, ACTIVE starts processing at once and delivers results to the contact points given.
get_monitorGet monitorGETGets one monitor's full configuration (objective, data schema, query, schedule, watchlist, contact point ids, status) by uuid from list_monitors or create_monitor.
list_monitorsList monitorsGETLists your organization's AI monitors with their full configuration. sortBy and sortOrder are required; archived monitors are left out unless status ARCHIVED is asked for.
update_monitorUpdate monitorPATCHChanges a monitor's configuration; only the fields sent change, and entityGroups and contactPointIds are replaced whole. Change whether it runs with activate_monitor, pause_monitor or archive_monitor.

Source group, 5 operations

create_source_groupCreate source groupPOSTCreates a private source group, a named set of 1 to 5,000 publisher domains usable by name in the sourceGroup filter of search_articles and monitor queries. Returns its numeric id.
delete_source_groupDelete source groupDELETEDeletes a source group your organization owns, by numeric id from list_source_groups.
get_source_groupGet source groupGETGets a source group your organization owns, with its domains, by numeric id from list_source_groups or create_source_group.
list_source_groupsList source groupsGETLists source groups (named sets of publisher domains) your organization owns plus public ones. sortBy and sortOrder are required. A group's name is used in the sourceGroup filter of search_articles.
update_source_groupUpdate source groupPATCHChanges a source group your organization owns (name, display name, description, domains); only the fields sent change.

Watchlist, 5 operations

create_watchlistCreate watchlistPOSTCreates a watchlist of up to 100 people and companies combined; people and companies are both required (send [] for none). Returns its numeric id for watchlistId of create_monitor.
delete_watchlistDelete watchlistDELETEDeletes a watchlist your organization owns, by numeric id. A watchlist attached to active monitors cannot be deleted.
get_watchlistGet watchlistGETGets a watchlist your organization owns, with its people and companies, by numeric id from list_watchlists or create_watchlist.
list_watchlistsList watchlistsGETLists watchlists (sets of people and companies) your organization owns plus public ones. sortBy and sortOrder are required. Ids go to watchlistId of create_monitor.
update_watchlistUpdate watchlistPATCHChanges a watchlist; only the fields sent change, and people or companies replace the whole list. Monitors linked to it use the new entities on later processing.

Article, 3 operations

peek_article_refreshPeek article refreshPOSTReturns refreshed article data already cached for up to 100 article ids, best effort, without scheduling any processing or creating a job.
summarize_articlesSummarize articlesPOSTWrites one AI summary of up to 100 articles matching the search_articles filters, guided by prompt; returns the summary and the articles used. method CLUSTERS takes one article per story.
vector_search_articlesVector search articlesPOSTSemantic search over news articles (the index holds about the last six months): send a natural-language prompt and get the closest articles with a relevance score. Without pubDateFrom only the last 30 days are searched.

Resolve, 2 operations

resolve_source_groupsResolve source groupsGETLooks up source groups by name (up to 100): for each name, your organization's private group if one exists, otherwise the public group of that name.
resolve_watchlistsResolve watchlistsGETLooks up watchlists by name (up to 100): for each name, your organization's private watchlist if one exists, otherwise the public watchlist of that name.

Activate, 1 operation

activate_monitorActivate monitorPOSTStarts a DRAFT or STOPPED monitor (status ACTIVE): it begins processing new content and delivering results to its contact points. Needs a valid configuration and plan capacity.

Article refresh job, 1 operation

submit_article_refresh_jobSubmit article refresh jobPOSTSubmits up to 100 article ids for background refresh and returns a jobId; read results with get_article_refresh_job until its status is COMPLETED. Fresh cached articles come back at once.

Event, 1 operation

list_monitor_eventsList monitor eventsGETLists the structured events an EVENT or MENTIONS monitor produced; sortBy and sortOrder are required. Each event's data holds the fields of the monitor's data schema.

History, 1 operation

list_story_historyList story historyGETLists previous generated versions of stories (by clusterId from search_stories) with their summaries, key points, questions and changelogs. page starts at 0.

Job, 1 operation

get_article_refresh_jobGet article refresh jobGETGets an article refresh job by jobId from submit_article_refresh_job: status PENDING, PROCESSING or COMPLETED, and the per-article results so far (each PENDING, COMPLETED or ERROR).

Journalist, 1 operation

get_journalistGet journalistGETGets one journalist's profile by id (from matchedAuthors or journalists in search_articles, or search_journalists). The email address is withheld.

Newsletter, 1 operation

list_monitor_newslettersList monitor newslettersGETLists the newsletters (digests of recent events) a monitor generated on its schedule, typically a TOPIC monitor; sortBy and sortOrder are required.

Pause, 1 operation

pause_monitorPause monitorPOSTStops an ACTIVE monitor (status STOPPED): it keeps its configuration and past output and delivers nothing new. Restart it with activate_monitor.

Stats, 1 operation

get_story_statsGet story statsGETCounts matching stories by creation time, grouped by splitBy (hour, day, week or month; default month), with the filters of search_stories.

Summary, 1 operation

list_monitor_summariesList monitor summariesGETLists the AI-generated summaries produced for a monitor as matching content is processed; sortBy and sortOrder are required.

Usage limit, 1 operation

get_usage_limitsGet usage limitsGETReturns your account's usage and limits: requests used and allowed in the current period, max page size, pagination limit, content truncation and reset date. Not counted against quota.

Wikipedia page, 1 operation

vector_search_wikipediaVector search wikipediaPOSTSemantic search over Wikipedia pages: send a natural-language prompt and get the closest page sections with a relevance score.

Use Perigon from Claude Code, Cursor or Codex

Connect Perigon in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Perigon operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Perigon

Create your workspace
Sign up free. Nothing to install.
Connect Perigon once
Add your Perigon connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.