Better tool selection
Help identify capabilities that fit the task, so agents can make more useful choices about their next action.
MCP Search
A feature of GoatFish
MCP Search
Find the right tools
for the task.
MCP Search is being developed within GoatFish to help connect AI tasks with relevant MCP tools. Its focus is better tool selection and routing, with more efficient use of model context and resources.
WITHIN THE GOATFISH WORKFLOW
A task to completeThe goal and its relevant context
MCP SearchFind relevant tool capabilities
GoatFish orchestrationCoordinate agents, tools, and work
Conceptual workflow. Availability and supported integrations will be defined through the beta.
The opportunity
Useful intelligence needs
relevant tools.
AI workflows can span many systems, each offering different capabilities. Choosing which tool fits a task is an important part of turning a request into useful work. MCP Search focuses on this discovery and selection step within GoatFish.
The Model Context Protocol (MCP) provides a common way for AI applications to connect to external tools and data. MCP Search applies that context to GoatFish’s broader goal: coordinating agents and tools around a shared plan.
Design goals
More relevance. Less overhead.
The beta will help evaluate these goals in practical workflows.
Focused context
Reduce unnecessary tool information in the working context, with the aim of lowering token overhead and associated cost.
A clearer path to action
Connect tool discovery with GoatFish’s orchestration of agents and workflows, keeping the user’s goal central.
Built into the bigger picture
One feature.
A broader workflow.
GoatFish brings agents and tools together to build software and complete complex work. MCP Search has a focused role within that experience: helping find the capabilities relevant to the work at hand.
Workflows to explore
Research and reporting. Identify relevant information-retrieval capabilities for a research task.
Software and small apps. Help builders find tools that support a development workflow.
Business automation. Explore how tool selection can support a workflow across connected systems.
Illustrative use cases. Specific connectors and workflow availability depend on the beta configuration.
Questions, answered
Understanding MCP Search.
How does MCP Search relate to GoatFish?
MCP Search is a feature of GoatFish. It focuses on finding relevant MCP tool capabilities within the broader agent-orchestration experience.
What does MCP mean?
MCP stands for Model Context Protocol, an open standard through which AI applications can connect to tools and data sources. MCP servers describe the capabilities they expose, allowing compatible applications to discover and use them.
Who is the beta intended for?
The developer launch is intended to gather feedback from novice and experienced builders, creators, and MCP practitioners. Contact AGI to discuss a suitable evaluation workflow and participation.
Which integrations and performance results are available?
Contact the AGI team for the current beta scope and supported integrations. Tool-selection quality, token usage, and cost will need to be evaluated for each workflow; results depend on the tools, models, and task involved.
Build with us
Bring a real workflow.
Help shape MCP Search within GoatFish. Tell us what you are building, which tools you use, and where tool discovery slows you down.