Base44
Build and edit Base44 apps from an AI assistant, and query data your app's users submit.
Create services, provision Postgres and Key Value stores, trigger deploys, query databases and read logs.
Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.
Paste it into any MCP client. Setup docs
The Render MCP server connects a Render cloud account to Claude, ChatGPT and any MCP-compatible agent. Its 25 tools create web services, static sites, cron jobs, Postgres databases and Key Value stores, trigger deploys, read logs and metrics, and run read-only SQL. Sign-in is OAuth or a Render API key.
Verified connector
Listed by Anthropic as a partner connector in its Connectors Directory.
Connection checked by Agentman on .
Anthropic states this reflects the level of review a connector received, not a security audit.
Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.
create_web_service cannot create Docker or container-registry services.update_* tools are no-ops that return a dashboard link.scopes_supported key.Ratelimit-Remaining and Ratelimit-Reset headers.select_workspace is deprecated and scheduled for removal.No. None of the 25 tools deletes, suspends or destroys anything. Render's REST API publishes Delete service, Delete Postgres instance, Suspend service and Restart service endpoints, and its MCP documentation states that for modifications and deletions beyond deploys and environment variables you must use the Render Dashboard or the REST API instead.
Four of the five create tools default to a free plan and one does not. Render's published server source defaults create_web_service, create_postgres and create_key_value to free, and static sites are free to deploy. Cron jobs have no free instance type, so create_cron_job defaults to Starter at 0.00016 dollars per minute of runtime.
No. The query_render_postgres tool runs SQL inside a read-only Postgres transaction, which Render states in the tool's own description and implements in the published server source by opening the transaction in read-only access mode. A write statement fails at the database rather than being filtered by the server.
Not through the MCP server. Render supports rollbacks and its REST API publishes a Roll back deploy endpoint, but no rollback tool exists among the 25. The recovery path is the Render Dashboard or the REST API. Render notes that a dashboard rollback also disables automatic deploys, while an API rollback does not.
It returns a dashboard link rather than changing anything. Render's server README states that direct updates to a web service are not supported through the MCP server and that the tool returns a link to the service's settings. The same is true of update_static_site and update_cron_job. Only update_environment_variables writes.
Not by default. Render's published server source documents that supplied variables merge with the service's existing ones, so the agent never needs to read your secrets into its context. Setting the replace parameter to true instead replaces the whole set, which deletes any variable the agent omitted.
None are published. We fetched Render's OAuth protected-resource descriptor and its authorization server metadata on 2026-08-18, and neither document carries a scopes_supported key at all. There is no narrower grant to choose, so the consent screen is the only place the connection's breadth is described.
Yes. Render's MCP documentation gives API key setup alongside OAuth for Claude Code, Claude Desktop, Codex and Cursor. Note that a Render API key carries your full REST API reach, including the delete and suspend endpoints the MCP tools do not expose, so an OAuth sign-in is the narrower of the two credentials.
Connect Render once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.