Capacity is not just a node count
Polygres graph capacity uses weighted units: active nodes plus active edges divided by ten. An application with a modest number of entities can still have a substantial relationship set, so estimate both parts of the model.
For example, 100,000 active nodes and 1,000,000 active edges represent 200,000 graph units. This is a capacity calculation, not a claim about the speed or complexity of any particular graph query.
Know what the base project includes
Free Nano includes 100,000 graph units alongside its storage and AI Context limits. Paid Basic starts at $16 per month and includes 200,000 graph units, 200,000 AI Context points, and 2 GiB of storage.
Additional graph capacity costs $1 per 100,000 units per month under the current Basic policy. The public calculator follows the dashboard’s rates, limits, and increments. Check the final account-specific quote in the dashboard.
Model relationships that serve a question
Adding every conceivable edge can make a graph harder to maintain and explain. Define the relationships that the application needs to traverse, including their direction, source, and lifecycle.
When the same underlying fact changes, decide which relationships must change with it. A clear update policy matters as much as the initial import because stale edges can connect otherwise current records in the wrong way.
Estimate with an actual sample
Count the nodes and edges generated by a representative slice of your source. Use that sample to understand how the model expands as more records are included.
Capacity is one planning input. Query behavior, source freshness, and the quality of the relationship model still need their own checks. A larger allowance cannot compensate for edges that do not represent the questions your application needs to answer.
Explore the next step
Continue with the project capacity calculator. For the current setup and API contract, use the Polygres documentation.