Loop
How it worksThe loopWhy LoopPricingDocs
compare Loop vs TomTom Places

Loop is built for agents.
TomTom is built for navigation.

If you’re building an AI agent that finds, checks, or acts on local business data, you need a data layer that speaks agent — structured output, freshness signals, confidence scores, and a feedback channel. That’s Loop. TomTom Places was designed for navigation UIs and geocoding, not language models.

feature comparison
FeatureLoopTomTom Places
Designed forAI agents (MCP + REST)Navigation UIs and geocoding
ProtocolMCP + RESTREST only
Output formatStructured JSON (typed schema per vertical)POI JSON optimised for routing, not reasoning
Freshness signalsobserved_at + confidence score on every recordNone
Availability labelingExplicit inferred: true until verifiedUnlabeled (looks authoritative)
Feedback loopreport() mutates record confidence and freshnessNone — read-only
Live verifyverify() re-checks a specific record on demandNone
Agent setup timeOne URL in any MCP client — no codeWrite wrapper code, parse responses, handle errors
Demand signalsEvery search leaves a signal — drives expansionBlack-box analytics only
CoverageRestaurants & salons, Kreuzberg Berlin (582 merchants; expanding by demand)Global
Free tierYes — no credit card, no applicationFree quota tier available, usage-based above
Open data attributionODbL (OpenStreetMap) + Apache 2.0 (Foursquare OS)Proprietary — TomTom ToS restrictions
Out-of-coverage responseExplicit honest error with suggested_actionEmpty results or low-confidence matches (silent failure)
Use Loop when
  • Building an AI agent (Claude, ChatGPT, custom)
  • You need structured, typed JSON output per vertical
  • Freshness and confidence matter for the user’s decision
  • You want to close the feedback loop with report()
  • You’re working in Kreuzberg Berlin
  • You want MCP access with no code setup
Use TomTom Places when
  • You need global coverage (Loop is Berlin-first)
  • You’re building a navigation or routing UI
  • You need geocoding or reverse geocoding
  • Your agent doesn’t need to report outcomes
common questions

When should I use Loop instead of TomTom Places?

Use Loop when you're building an AI agent that needs to search, verify, or act on local business data. Loop speaks MCP natively, returns typed JSON with freshness signals, and has a report() action that closes the feedback loop. Use TomTom Places when you need global geocoding or routing context for a navigation application.

When should I use TomTom Places instead of Loop?

Use TomTom Places when global coverage is the primary requirement, or when building for navigation and routing use cases. Loop currently covers restaurants and salons in Kreuzberg Berlin, with more verticals and areas expanding. If your users are outside that area or if you need turn-by-turn routing data, TomTom may be a better fit for now.

Does Loop replace TomTom Places in an agent pipeline?

For the covered area (restaurants and salons in Kreuzberg Berlin), yes — Loop gives agents better-structured data with freshness signals, MCP access, and a feedback channel. For out-of-coverage queries, Loop returns an explicit structured error so agents can fall back gracefully.

Start building with Loop.

Add https://stayinloop.dev/mcp to your MCP client, or call the REST API directly. Free tier, no credit card.