AI coding assistant with direct access to your Convex backend
Visit WebsiteFreemiumVisit Website
Tracked since2026
What is Convex MCP?
Convex is a reactive backend platform that bundles a document database, serverless functions, scheduling and file storage behind a single TypeScript codebase. Its MCP server ships with the Convex CLI and runs against a chosen deployment, which gives a coding assistant a direct view of the backend it is being asked to write code for. The server exposes deployment status, the list of tables with both declared and inferred schemas, paginated document reads from any table, and the function spec listing every query, mutation and action with its argument types. An agent can call a deployed function with arguments and see the real response, or run a sandboxed read-only JavaScript query for exploration without deploying anything first. Environment variables can be listed, read and set, and deployment logs are readable for debugging. This is aimed at teams building on Convex with an assistant in the editor. Instead of inferring the data model from source files, the assistant reads the actual schema, samples a few documents, checks what a function returns, then writes code that matches. Function calls can mutate data and environment variables can be rewritten, so point the server at a development deployment before letting it anywhere near production.
Preview
Key Features
Lists deployment tables with both declared and inferred schemasReads documents from any table with paginationReturns the function spec for queries, mutations and actions with argument typesExecutes deployed functions with arguments and returns the real responseRuns sandboxed read-only JavaScript queries without deploying codeLists, reads and sets deployment environment variables
Pricing
Freemium
Convex MCP offers a generous free tier with optional paid upgrades for advanced features.
Reviews

$99Free with your review
Write a reviewReview Convex MCP, get a free AI guide
Share your experience and we will send you Improve Your Thinking Patterns Using ChatGPT, free.
Best Convex MCP Alternatives
Top alternatives based on features, pricing, and user needs.
SupabaseFreemium
Open source Firebase alternative
AWS AmplifyFreemium
Build and deploy full-stack web & mobile apps on AWS
PocketbaseFree
Open-source backend in a single file
AppwriteFreemium
Open-source backend for web, mobile, and Flutter. Self-host or cloud
NhostFreemium
Open-source Firebase alternative
Firebase MCPFreemium
Connect AI coding assistants to your Firebase backend
Still deciding?
Most buyers shortlist 2 or 3 tools before committing. Pull a side-by-side comparison or browse the full alternatives shortlist below.
Explore More
Convex MCP FAQ
How does Convex MCP help an AI assistant write code that matches the actual backend?
Convex MCP gives the coding assistant direct access to a live deployment, so it can read the actual schema, sample documents from any table, and inspect function signatures and return values. This means the assistant writes code against the real data model and behavior instead of guessing from source files.
Which teams benefit most from using Convex MCP?
Teams building on Convex with an AI assistant in their editor benefit most from Convex MCP. It is designed for developers who want the assistant to understand the exact state of their backend, including tables, functions, and environment variables, to generate accurate code.
How does Convex MCP compare to Firebase MCP for backend-aware coding assistance?
Convex MCP is built specifically for the Convex platform, which bundles a document database, serverless functions, scheduling, and file storage in a single TypeScript codebase. Unlike Firebase MCP, Convex MCP exposes both declared and inferred schemas, allows sandboxed read-only queries, and can call deployed functions with real arguments to verify responses.
What are the main limitations or trade-offs of using Convex MCP?
Convex MCP can mutate data and rewrite environment variables when it calls functions, so it must be pointed at a development deployment before being used near production. Additionally, it only works with Convex backends, so teams using other platforms cannot benefit from its direct backend integration.
How is Convex MCP priced?
Convex MCP is available on a free tier, with paid plans for more usage and features. The pricing model scales with the level of activity and access needed.
Can Convex MCP read deployment logs for debugging?
Yes, Convex MCP can read deployment logs to help debug issues. This allows the AI assistant to inspect runtime behavior and errors directly from the backend.
Does Convex MCP allow an AI assistant to run sandboxed queries without deploying changes?
Yes, Convex MCP supports running sandboxed read-only JavaScript queries for exploration without deploying anything first. This lets the assistant test queries safely before writing production code.
Which features does Convex MCP expose to an AI assistant for backend inspection?
Convex MCP exposes deployment status, the list of tables with schemas, paginated document reads, the function spec with argument types, environment variables, and deployment logs. It also allows calling deployed functions with arguments and running sandboxed queries.
Source: convex.dev