Pointmoon documentation
Read the conditions at a place.
Pointmoon combines weather, daylight, terrain and nearby observations in a structured response. Your application decides how to use the readings.
Make a request
Use a public location. The facts endpoint is read-only and currently needs no account or API key.
curl "https://pointmoon.ai/api/moon?audience=facts&surface=open&lat=42.36&lng=-71.06"Supply both latitude and longitude in decimal degrees. When you only have a name, use place instead:
curl "https://pointmoon.ai/api/moon?audience=facts&surface=open&place=Boston"Coordinates take precedence when both forms are supplied. Use a specific location and inspect the resolved place before presenting a local claim.
Download the OpenAPI specification for the public facts request. It describes this supported entry point, not private Studio APIs.
Handle the response
A successful request returns JSON. Read the schema version before consuming it. The facts contain sourced signals and a field snapshot; optional readings can be absent or unresolved.
- Keep observations, forecasts and inferences distinct.
- Check each reading’s source and freshness. Nearby wildlife observations do not promise what someone will see today.
- When evidence is missing, show that uncertainty. Do not turn an unresolved reading into a claim.
- Handle non-success HTTP statuses separately. Respect rate limits and back off before retrying.
The response contract describes the consumer boundary. Coverage varies by source and location.
Connect an agent
The public Streamable HTTP MCP server is https://pointmoon.ai/api/mcp. Its tool listing supplies the input schemas for field_truth and step_outside.
Access and licensing
Free for non-commercial use, subject to fair-use limits and source attribution. Commercial use needs a paid plan, by request. The Pointmoon engine is proprietary; the separate MCP connector is open source. No uptime guarantee applies to the free service.