Vectoree

Vectoree covers what a coding agent still needs after it has written an app: model access, login, a database and file storage, which usually means four services and four keys. It puts them in one console. Existing code keeps using the OpenAI SDK; only the base URL and the key change.

Highlights

• Model Plaza: one key for models from providers including OpenAI, Anthropic, Google, xAI and DeepSeek, with a default vectoree/auto route to a cheaper available model
• Tool Hub: search, Similarweb, Semrush and Ahrefs data over a hosted MCP server, using the same project key
• Email and Google sign-in for your app's users, one Postgres database per project with row-level access control, and public or private storage buckets
• An agent skill plus the @vectoree/cli package, so the agent runs single-step commands such as login, link, db and storage
• Starter templates whose prompts an agent can paste and build on
• A one-command setup that points Codex CLI at the Vectoree gateway

Good to know

• Billing is prepaid: an empty wallet stops gateway and tool calls, and per-million-token prices are listed in the dashboard
• Hosting on a public URL (Deploy) and charging your own users (Payments) are announced but not available yet
• Edge functions, realtime and compute are not part of the current release

For solo developers and small teams who build with coding agents and want the backend wired up in the same session.