Open source MCP bridge

One bridge. 2,000+ models.

Discover, compare, run, and track AI models from the clients you already use.

Install the bridge Star on GitHub

SandBase CLI connecting a command line interface to image, audio, data, search, and video AI services

A complete model workflow

From question to result, without leaving your agent.

Discover

Search a catalog of more than 2,000 models and APIs by task, provider, or capability.

Inspect

Read schemas, pricing, and execution requirements before spending credits.

Run

Execute text, image, video, audio, and multimodal workloads through one interface.

Track

Poll asynchronous jobs and review recent cost and balance from the same client.

One command setup

Connect the tools you already use.

The interactive setup writes the right MCP configuration for your selected client.

$ brew install sandbaseai/tap/sandbaseai-cli
$ sandbase connect

$ npx -y https://github.com/sandbaseai/cli/releases/download/v0.1.17/sandbaseai-cli-0.1.17.tgz connect

Six MCP tools

A compact interface for a large model catalog.

Discover first. Inspect before you run.

sandbase_discover

sandbase_inspect

Inputs, outputs, pricing, and requirements.

sandbase_run

Execute a selected model or API.

sandbase_run_get

Follow asynchronous work to completion.

sandbase_runs

Review recent runs and cost.

sandbase_account

Check account balance and status.

Verify the project

Public code. Public registry.

Apache-2.0 licensed and independently discoverable across the MCP ecosystem.

Official MCP Registry Official Homebrew tap Source and releases Registry showcase Independent directory profile

Give your agent a wider model toolbox.

Install SandBase CLI and choose your client.

Start with SandBase CLI