FlowRunner
PricingContact
Theme
Start Free

Give AI agents location intelligence from the Foursquare Places API. Agents search places by text or coordinates, pull rich venue details like hours, ratings, and price tier, run type-ahead autocomplete, and surface user photos and tips for any venue.

6 actions API key available
Foursquare website ↗ Platform Documentation ↗ Capability data verified 2026-07-28
A new business location arrives in the intake form as a name and rough address
Search Places matches the submission against Foursquare's venue database
Get Place Details pulls the canonical record: address, coordinates, contact, website, hours, rating, price tier
Submissions that match multiple candidate venues, or none, are separated from the clean matches
Get Place Tips and Get Place Photos attach crowd-sourced context to the record
Enriched records sync to the CRM and the review queue posts to the team channel
An operator resolves the ambiguous matches, picking the right venue or marking the submission unmatched, because writing the wrong canonical place ID onto a record poisons every workflow that reads it later

What This Integration Enables

The gap between "a place someone typed" and "a place that verifiably exists, with hours, contact, and a track record" is where location-dependent operations lose time. Foursquare's Places API closes that gap with one of the most mature venue databases in the industry, and this connector makes it an agent capability. FlowRunner agents search places by text or coordinates, hydrate full venue details, run type-ahead autocomplete, list what is physically nearby, and surface user photos and tips for any venue. - Resolve submitted names and addresses to canonical place IDs with categories and coordinates - Pull venue details: contact, website, hours, rating, price tier, popularity - Build "what's around this point" lists ordered by distance - Power type-ahead search in intake forms and chat flows - Attach crowd-sourced tips and photos to venue records The connector reads a third-party database; nothing here destroys data or spends beyond API usage. The judgment that needs a human, and gets one through FlowRunner's [human-in-the-loop](/concepts/human-in-the-loop) steps, is identity: which real-world place a record actually refers to.

Without FlowRunner

Venue data is typed and trusted Whatever the form said becomes the record: misspelled names, dead addresses, no hours
"What's near this location" is a browser task Someone opens a map, squints, and types what they see into a note
Venue quality is unknowable at scale Rating, price tier, and whether the place still exists require a manual check each

With FlowRunner

Submissions resolve to canonical places Names and addresses match to Foursquare place IDs with structured detail attached
Nearby context is a workflow step Distance-ordered venue lists generate automatically around any coordinate
Crowd signal enriches every record Ratings, tips, and photos attach to venues before decisions are made about them

Use Case Scenarios

Location intake that produces canonical records

Field teams submit new client sites by name and rough address. The agent runs Search Places, hydrates the best match with Get Place Details, and writes the structured venue, address, coordinates, phone, hours, into [HubSpot](/integrations/hubspot). One-match submissions flow through; multi-match and no-match cases queue for a person with the candidates side by side. The CRM stops accumulating venues that exist only as typos.

Site research assembled before the visit

Before a market visit, the agent takes each target coordinate, pulls Get Nearby Places for the surrounding retail landscape, and hydrates the notable venues with details, tips, and photos. The briefing lands in [Google Sheets](/integrations/google-sheets) with a summary per site posted to [Slack](/integrations/slack). The person walking the market arrives already knowing the block: what is there, what it charges, and what customers say about it.

Partner vetting with crowd evidence

When a prospective venue partner enters the pipeline, the agent pulls Get Place Details for operating status, rating, and price tier, then Get Place Tips for the unfiltered customer voice. The digest attaches to the deal record. A partnership manager reads three years of "great space, chaotic service" tips before the first call, and the venues that look better in their pitch deck than in their tips get the scrutiny they earned.

Human-in-Loop Highlight

The consequential act in this connector is not an API call, it is the binding: writing an fsq_place_id onto a customer, site, or partner record declares "this record means that physical place," and every downstream workflow, routing, outreach, reporting, inherits the declaration without questioning it. Search Places returns ranked candidates, and ranked means the top hit is sometimes the similarly named cafe two towns over. So the match confidence drives the gate. Single clean matches bind automatically; anything ambiguous, multiple plausible candidates, a no-match on a submission that should exist, a category that contradicts the record, goes to an operator who sees the submission and the candidates with their addresses and categories side by side. Correcting a wrong binding after months of workflows have acted on it is archaeology. Preventing it is one human minute at exactly the right moment.

Agent processes routinely
Detects exception requiring judgment
Clear match Continues automatically
Ambiguous Routes to human via preferred channel
Human decides
Agent resumes with decision

Agent Capabilities

6 actions

Places

3
  • Search Places Searches points of interest by text query and optional lat,lng with radius, category filters, and result count. Returns ranked places with place ID, name, categories, location, and coordinates.
  • Get Place Details Retrieves a venue's full record by place ID: name, categories, address, coordinates, contact, website, hours, rating, price tier, and popularity when available.
  • Get Nearby Places Returns venues physically near a coordinate, ordered by distance in meters. The read behind "what's around this point" briefings.

Search

1
  • Autocomplete Returns real-time suggestions for a partial query biased toward a location, covering places, addresses, regions, and categories, each with a place ID ready for detail hydration.

Content

2
  • Get Place Photos Retrieves user-contributed photos for a venue, sortable by newest or most popular, with prefix and suffix parts that assemble into full image URLs at any size.
  • Get Place Tips Retrieves user-contributed tips for a venue with text, date, and agreement counts, sortable by newest or most popular. The crowd-sourced signal behind vetting workflows.

Frequently Asked Questions

What can FlowRunner do with Foursquare?

FlowRunner agents can run Search Places, Get Place Details, and Autocomplete in Foursquare, plus 3 more actions.

Does connecting Foursquare to FlowRunner require OAuth?

No. Foursquare connects to FlowRunner with an API key, no OAuth flow required.

Can Foursquare trigger a FlowRunner workflow automatically?

Foursquare doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with Foursquare

$100 in credits. No card required. Connect in minutes.