Tavily is a powerful API connecting AI agents to the live web for real-time search, extraction, research, and web crawling. It provides a production-grade retrieval stack to ground LLMs with fresh, factual web context, reducing hallucinations.
Built for scale, Tavily handles millions of requests with low latency and built-in safeguards against PII leakage and prompt injection. Trusted by over one million developers and major enterprises like MongoDB and IBM, it offers seamless integration with leading LLM providers for sophisticated AI applications.
Built for scale, Tavily handles millions of requests with low latency and built-in safeguards against PII leakage and prompt injection. Trusted by over one million developers and major enterprises like MongoDB and IBM, it offers seamless integration with leading LLM providers for sophisticated AI applications.
OpenAI has expanded its Responses API to facilitate the development of agentic workflows. This includes support for a shell tool, an agent execution loop, a hosted container workspace, context compaction, and reusable agent skills. The new features aim to offload the complexities of building execution environments from developers, providing a managed infrastructure for handling tasks like file management, prompt optimization, secure network access, and handling timeouts.
A core component is the agent execution loop, where the model proposes actions (running commands, querying data) that are executed in a controlled environment, with the results fed back to refine the process. Skills allow for the creation of reusable task patterns.
A core component is the agent execution loop, where the model proposes actions (running commands, querying data) that are executed in a controlled environment, with the results fed back to refine the process. Skills allow for the creation of reusable task patterns.
Developers are replacing bloated MCP servers with Markdown skill files — cutting token costs by 100x. This article explores a two-layer architecture emerging in production AI systems, separating knowledge from execution. It details how skills (Markdown files) encode stable knowledge, while MCP servers handle runtime API interactions. The piece advocates for a layered approach to optimize context window usage, reduce costs, and improve agent reasoning by prioritizing knowledge representation in a version-controlled, accessible format.
OpenSandbox is a general-purpose sandbox platform for AI applications, offering multi-language SDKs, unified sandbox APIs, and Docker/Kubernetes runtimes for scenarios like Coding Agents, GUI Agents, Agent Evaluation, AI Code Execution, and RL Training.
This post breaks down why MCP servers fail, six best practices for building ones that work, and how Skills and MCP complement each other. It emphasizes designing MCP servers as user interfaces for AI agents, focusing on outcomes, flattened arguments, clear instructions, curation, discoverable naming, and pagination.
* **Focus on Outcomes, Not Operations:** Instead of exposing granular API endpoints as tools, create high-level tools that deliver the *result* the agent needs.
* **Flatten Arguments:** Use simple, typed arguments instead of complex nested structures.
* **Instructions are Context:** Leverage docstrings and error messages to provide clear guidance to the agent.
* **Curate Ruthlessly:** Limit the number of tools exposed and focus on essential functionality.
* **Name Tools for Discovery:** Use a consistent naming convention (service_action_resource) to improve discoverability.
* **Paginate Large Results:** Avoid overwhelming the agent with large datasets; use pagination with metadata.
* **Focus on Outcomes, Not Operations:** Instead of exposing granular API endpoints as tools, create high-level tools that deliver the *result* the agent needs.
* **Flatten Arguments:** Use simple, typed arguments instead of complex nested structures.
* **Instructions are Context:** Leverage docstrings and error messages to provide clear guidance to the agent.
* **Curate Ruthlessly:** Limit the number of tools exposed and focus on essential functionality.
* **Name Tools for Discovery:** Use a consistent naming convention (service_action_resource) to improve discoverability.
* **Paginate Large Results:** Avoid overwhelming the agent with large datasets; use pagination with metadata.
This article compares Model Context Protocol (MCP), Function Calling, and OpenAPI Tools for integrating tools and resources with language models, outlining their strengths, limits, security considerations, and ideal use cases.
Interact with opencode server over HTTP. The `opencode serve` command runs a headless HTTP server that exposes an OpenAPI endpoint that an opencode client can use.
The Universal Tool Calling Protocol (UTCP) is an open standard that describes how to call existing tools directly, eliminating the need for wrappers. It focuses on direct communication with tool endpoints (HTTP, gRPC, WebSocket, CLI, etc.) to reduce latency and maintain existing security and billing systems.
This article lists and ranks the top Model Context Protocol (MCP) servers on GitHub as of June 2025, highlighting their capabilities and emphasizing the importance of security when granting agents access to sensitive data. It positions Pomerium as a solution for enforcing policy and securing agentic access to MCP servers.
|**GitHub Repository** |**Description** |
|---------------------------------|-----------------------------------------------------------------------------|
| github/github-mcp-server | Manages GitHub issues, pull requests, discussions with identity & permissions. |
| microsoft/playwright-mcp | Triggers browser automation tasks (QA, scraping, testing). |
| awslabs/mcp | Exposes AWS documentation, billing data, and service metadata. |
| hashicorp/terraform-mcp-server | Secure access to Terraform providers and modules. |
| dbt-labs/dbt-mcp | Exposes dbt’s semantic layer and CLI commands. |
| getsentry/sentry-mcp | Access to Sentry error tracking and performance telemetry. |
| mongodb-js/mongodb-mcp-server | Interacts with MongoDB and Atlas instances securely. |
| StarRocks/mcp-server-starrocks | Brings MCP to the StarRocks SQL engine. |
| vantage-sh/vantage-mcp-server |Focuses on cloud cost visibility. |
|**GitHub Repository** |**Description** |
|---------------------------------|-----------------------------------------------------------------------------|
| github/github-mcp-server | Manages GitHub issues, pull requests, discussions with identity & permissions. |
| microsoft/playwright-mcp | Triggers browser automation tasks (QA, scraping, testing). |
| awslabs/mcp | Exposes AWS documentation, billing data, and service metadata. |
| hashicorp/terraform-mcp-server | Secure access to Terraform providers and modules. |
| dbt-labs/dbt-mcp | Exposes dbt’s semantic layer and CLI commands. |
| getsentry/sentry-mcp | Access to Sentry error tracking and performance telemetry. |
| mongodb-js/mongodb-mcp-server | Interacts with MongoDB and Atlas instances securely. |
| StarRocks/mcp-server-starrocks | Brings MCP to the StarRocks SQL engine. |
| vantage-sh/vantage-mcp-server |Focuses on cloud cost visibility. |
This tutorial details how to use FastAPI-MCP to convert a FastAPI endpoint (fetching US National Park alerts) into an MCP-compatible server. It covers environment setup, app creation, testing, and MCP server implementation with Cursor IDE.
