# SmartGate > SmartGate is an MCP-native algorithm gateway. It sits between AI agents and the > open web: it fetches, searches, compresses, deduplicates, and hard-caps tool > traffic before it reaches your model provider. It is not an LLM proxy — you keep > your own host model (Cursor, Claude, OpenClaw, …) and connect SmartGate as an > MCP tool server over Streamable HTTP with Bearer auth. ## Documentation - [Overview](https://smartgate.network/docs): What SmartGate is, the seven MCP tools, and plans at a glance. - [Quick Start](https://smartgate.network/docs/installation): Create an account, mint an API key, paste the endpoint into your MCP client, verify the connection. - [Connect Guide](https://smartgate.network/docs/connect): Per-host setup for Cursor, Claude, Windsurf, OpenClaw, and generic SSE clients. - [MCP Endpoint](https://smartgate.network/docs/mcp-endpoint): Endpoint URL, authentication, per-host configuration, and REST audit headers. - [Tools Reference](https://smartgate.network/docs/tools): Parameters and behaviour of smart_fetch, smart_search, smart_context_gate, smart_dedup, smart_budget_guard, smart_memory, and smart_pipe. - [Activity Logs & Reports](https://smartgate.network/docs/logs): Team-scoped audit records, trace links, and usage reports. - [Security](https://smartgate.network/docs/security): API key hygiene, team isolation, log retention, and rate limits. - [Billing](https://smartgate.network/docs/billing): Plans, Paddle checkout, upgrades, and subscription management. - [Teams & Invitations](https://smartgate.network/docs/team): Team roles (OWNER/ADMIN/MEMBER), invitations, and plan limits. - [Webhooks (Preview)](https://smartgate.network/docs/webhooks): Register outbound webhook endpoints — Preview: endpoints and signing secrets are managed, business events are not delivered yet. ## Product - [Token Control](https://smartgate.network/features/token-control): Compress, deduplicate, and budget-cap tool traffic before it bills your model. - [Traffic Customs](https://smartgate.network/features/traffic-customs): Fetch, search, and pipeline shaping for agent traffic. - [Audit & Compliance](https://smartgate.network/features/audit-compliance): Per-tool audit records, trace linking, and export. - [Pricing & Plans](https://smartgate.network/pricing): Free, Pro, Teams, and Enterprise tiers with token caps and rate limits. - [Solutions by Role](https://smartgate.network/solutions): Role-based pages for the people who own AI spend, risk, and delivery. - [Industries](https://smartgate.network/industries): Governed AI traffic guidance by sector. - [Integrations](https://smartgate.network/integrations): Connecting SmartGate to observability, auth, and AI infrastructure. ## Blog - [SmartGate Blog](https://smartgate.network/blog): Releases, MCP notes, and agent cost engineering. - [News](https://smartgate.network/blog/category/news): Updates, releases, and announcements. - [Education](https://smartgate.network/blog/category/education): Educational content on MCP, routing, and cost optimization. ## Contact - [Contact](https://smartgate.network/contact): Billing, technical, and partnership questions. ## Optional - [Privacy Policy](https://smartgate.network/privacy) - [Terms of Service](https://smartgate.network/terms) - [GitHub](https://github.com/smartgate-network) ## Articles Full articles published on smartgate.network. Each one answers a specific question about governing MCP tool traffic; the machine-readable bundle of all of them is at https://smartgate.network/llms-full.txt. ### Integrations - [AI Agent Architecture: LangGraph vs CrewAI vs a Gateway](https://smartgate.network/integration/agent-infra-stack-comparison-langgraph-vs-smartgate): An AI agent architecture is drawn as four layers and decided as a framework choice. The twelve shipped symbols here cover what neither LangGraph nor CrewAI decides: registration. - [AI Automation for SaaS Operations: Queues, Webhooks, Billing](https://smartgate.network/integration/ai-automation-for-saas-operations): AI workflow automation in a SaaS product is three primitives, not a canvas: a durable email queue, a webhook intake that verifies before it deduplicates. - [AI Financial Analysis for Agents: From Audit Log to Report](https://smartgate.network/integration/ai-agents-for-financial-data-analysis): AI financial analysis with agents is an accounting problem before it is a model problem. Seven shipped functions price a call against a team, a window and a budget. - [AI Gateway vs API Gateway: Where Each Responsibility Sits](https://smartgate.network/integration/ai-gateway-vs-api-gateway): An AI gateway and an API gateway share routing, auth and request rate limits; they part at per-key token quota, cost accounting, the audit row per task and tool proxying. - [API Documentation Best Practices for AI Tools: Generators, Specs](https://smartgate.network/integration/api-documentation-best-practices-for-ai-tools): Twelve API documentation best practices for AI tools, each with code: publish a machine-readable spec, give agents annotations, test the docs against the code, enforce a frontmatter contract. - [Build a Low-Cost AI Backend Architecture: Count, Cache, Meter](https://smartgate.network/integration/build-low-cost-ai-backend-architecture): A low-cost AI backend is four cost decisions, not a cheaper model: price each call, cache and batch what repeats, route on cost per resolved task, and self-host only past a computed break-even. - [Concurrency Control in AI Backend Systems: Limits, Queues, Pools](https://smartgate.network/integration/concurrency-control-in-ai-backend-systems): Concurrency control in an AI backend has five layers: a plan catalog that decides the numbers, a two-level MCP limiter per key and per team, an edge sliding window for bursts. - [Connect OpenClaw to SmartGate over MCP: Config and Guardrails](https://smartgate.network/integration/openclaw): OpenClaw MCP setup in four steps: point OpenClaw at a Streamable HTTP MCP endpoint, send the Bearer plus platform headers, then cap tool traffic with per-plan rate limits, a hard budget guard. - [OpenRouter Alternative Explained: SmartGate vs OpenRouter](https://smartgate.network/integration/smartgate-vs-openrouter-differences): Looking for an OpenRouter alternative? Check the layer first: OpenRouter routes which model answers; SmartGate governs the tool calls that feed it, with compression. - [Optimize AI Agent Execution Cost: Meter, Attribute, Prove](https://smartgate.network/integration/optimize-ai-agent-execution-cost): Optimising AI agent execution cost starts with attribution, not model choice: record usage at the gateway, infer tokens when a tool omits them, count with the right tokeniser. - [Prompt Security for AI Applications: Policy, Keys, Audit](https://smartgate.network/integration/secure-prompt-handling-in-ai-applications): Prompt security is five layers, not one filter: a schema-validated policy per team, a clamp that only tightens, a key check on every request, masking before the log is written. - [Token Optimization Techniques for AI Apps: Caps and Cost Control](https://smartgate.network/integration/token-optimization-techniques-for-ai-apps): Token optimization for AI apps: measure usage, shape prompts, cache repeats, route by price, cap output. Eight techniques with the gateway code behind them. ### Guides - [Advanced RAG Architecture for AI Agents: Four Stages to Own](https://smartgate.network/industry/advanced-rag-architecture-for-ai-agents): An advanced RAG architecture for AI agents is four stages you own end to end: a rebuildable index, dedup before and after retrieval, compression with a reported ratio, and a search path with a floor. - [Agent Memory Architecture: What to Store and Retrieve](https://smartgate.network/industry/agent-memory-architecture): Agent memory architecture is a write path, a read path and a bill. What gets stored, when retrieval is allowed to run, and how a gateway exposes it as one tool. - [Agentic RAG Survey: How the Research Splits the Field](https://smartgate.network/industry/agentic-rag-survey): An agentic RAG survey in practice: the taxonomies the papers use, where search-as-a-tool, self-reflection and evaluation benchmarks diverge, and what to take from each. - [Agentic Search vs RAG: A Tool Call or an Index You Own](https://smartgate.network/industry/agentic-search-vs-rag): An agent's search tool reads the live web and someone else owns the ranking; a RAG index reads a corpus you built. Freshness, cost per call, and how to pick the path. - [Anthropic MCP: Claude Clients, Servers, and the Gateway Between](https://smartgate.network/industry/anthropic-mcp): Anthropic MCP is the Claude side of the Model Context Protocol: which clients speak it, how the transport moved to Streamable HTTP, and what a gateway adds in front. - [Anthropic Model Context Protocol: Names, Clients and Spec Work](https://smartgate.network/industry/anthropic-model-context-protocol): Anthropic's implementation of the Model Context Protocol, seen through the names its tooling uses: package scopes, console surfaces, the budget a request carries, and the spec pages it published. - [Build an MCP Client for AI Agents: Config, Auth, Transport](https://smartgate.network/industry/build-mcp-compatible-client-for-ai-agents): How to make an MCP client for AI agents actually connect: the per-platform config shapes each client demands, the two headers a gateway reads, the single upstream URL. - [Context Window Management Techniques for AI Agents](https://smartgate.network/industry/context-window-management-techniques): Context window management for agents in twelve real code paths: compression ratios, segmenting long input, pipeline truncation, semantic dedup with MMR reranking, team memory. - [Enterprise AI Gateway Architecture Best Practices: MCP Gateway](https://smartgate.network/industry/enterprise-ai-gateway-architecture-best-practices): An enterprise AI gateway is enforced by the layer around it: tenancy, per-request identity, canonicalized inputs, runtime config, scoped exclusions, validated parameters, and typed plan gates. - [How to Deploy an AI Gateway in a Private Cloud: MCP and Audit](https://smartgate.network/industry/how-to-deploy-ai-gateway-in-private-cloud): Deploy an AI gateway in your own cloud: the MCP endpoint you expose, the base URL that must fail loudly, the server list you version, and the audit window you operate. - [How to Enforce a Token Quota Per Team in an AI Gateway](https://smartgate.network/industry/how-to-enforce-token-quota-per-team): A per-team token quota takes four shipped pieces: the entitlement read, a pre-flight counter, a refusal your caller honours, and an alert at 80 and 100 percent. - [MCP Gateway: What It Adds Over a Raw MCP Server](https://smartgate.network/industry/mcp-gateway): An MCP gateway puts one upstream URL in front of every agent host, then adds auth forwarding, per-key and per-team limits, plan-clamped budget caps and an audit row per tool call. - [MCP Inspector Alternatives: Where to Debug a Live Server](https://smartgate.network/industry/mcp-inspector-alternatives): The MCP Inspector is one of five places to look when a tool call fails: a session trace, server logs, an audit row, a health check and the client's own pane. How to pick. - [MCP Logging and Observability: Audit Rows, Retention](https://smartgate.network/industry/mcp-logging-and-observability): MCP logging means one audit row per tool call. What the gateway records, what it masks, how long each plan keeps it, and what those rows cannot tell you. - [MCP OAuth and Authorization: Key Hashes, Bearer, Roles](https://smartgate.network/industry/mcp-oauth-and-auth): MCP OAuth defines the shape of an authorised request; the gateway owns the decisions. How a key is issued, hashed, validated, scoped to a team and capped. - [MCP Protocol Versions and Transports: HTTP, SSE, and stdio](https://smartgate.network/industry/mcp-protocol-versions-and-transports): How the MCP protocol negotiates versions and capabilities, which transport to use — Streamable HTTP, legacy SSE, or stdio — and what a gateway has to patch for older clients. - [MCP Resources, Prompts and Sampling Through a Gateway](https://smartgate.network/industry/mcp-resources-prompts-sampling): MCP resources, prompts and sampling sit beside tools. What a gateway can serve, meter and govern for each primitive, and the one direction it cannot bill. - [MCP Server Example: One Tool, from Config to First Call](https://smartgate.network/industry/mcp-server-example): A worked MCP server example end to end: the config entry a host reads, the tool list the server publishes, the first call, the record that proves it arrived, and the read path behind it. - [MCP Server Explained: Endpoint, Tool Registry, Sessions](https://smartgate.network/industry/mcp-server): An MCP server answers initialize, tools/list, and tools/call at one Streamable HTTP endpoint. What an MCP server is, when a session matters, and how to run one. - [MCP Server List: A Buyer's Checklist Before You Install](https://smartgate.network/industry/mcp-server-list): A buyer's checklist for evaluating an MCP server before you install it: the auth model, the tool surface, the limits, the audit trail and where the counters live. - [MCP Specification Walkthrough: How to Read the Spec](https://smartgate.network/industry/mcp-specification-walkthrough): A practitioner's walkthrough of the MCP specification: which chapters are normative, how initialize negotiates a version, the error shapes nobody defines, and the limits. - [MCP Tools Reference: tools/list Schemas and Annotations](https://smartgate.network/industry/mcp-tools-reference): A working reference for the MCP tools list: the schema and annotations tools/list returns, what read-only hints mean for hosts, and how a gateway normalizes both before they reach a client. - [MCP vs A2A Protocol: Vertical Tools, Horizontal Agents](https://smartgate.network/industry/mcp-vs-a2a-protocol): MCP wires one agent to tools; A2A wires agents to each other. What each protocol carries, where the two boundaries meet, and where a gateway sits when both are in play. - [MCP vs REST API: Choosing the Right Agent Interface](https://smartgate.network/industry/mcp-vs-rest-api): MCP and REST/OpenAPI answer different questions for agent tool access. Compare discovery, sessions, credentials and limits, and see what a gateway does for either surface. - [Model Context Protocol (MCP) Message Format Explained](https://smartgate.network/industry/mcp-protocol-message-format-explained): A Model Context Protocol message is JSON-RPC 2.0 with three shapes — requests, notifications, and results. Here is how initialize, tools/list, and tools/call actually travel over Streamable HTTP. - [Model Context Protocol Documentation: Spec, Schema, Changelog](https://smartgate.network/industry/model-context-protocol-documentation): Model Context Protocol documentation is versioned by date: a dated specification revision, a schema published beside it and a changelog of what moved. How to read each one. - [Model Context Protocol Example: A JSON-RPC Round Trip](https://smartgate.network/industry/model-context-protocol-example): A Model Context Protocol example, message by message: the initialize handshake, a tools/call round trip, and the JSON-RPC error body a client has to read. - [Model Context Protocol Explained: Actors, Lifecycle, Transport](https://smartgate.network/industry/model-context-protocol-explained): The Model Context Protocol end to end: its three actors, the initialize-and-list lifecycle, the two transports, and the capabilities a client and a server each declare. - [Model Context Protocol MCP: What the Acronym Maps To](https://smartgate.network/industry/model-context-protocol-mcp): Model Context Protocol MCP is one name for two halves: a client that carries context, a server that holds tools, and the mcp prefix that configuration keys and tool names carry. - [Model Context Protocol Server: Transport and Hosting](https://smartgate.network/industry/model-context-protocol-server): A Model Context Protocol server is hosted rather than spawned: pick a transport, publish one endpoint, decide what survives a restart, and plan the deployment shape before sharing it. - [Python MCP Server Tutorial: Build It, Then Govern It](https://smartgate.network/industry/python-mcp-server-tutorial): Build an MCP server in Python with the official SDK: one tool, one resource, a local smoke test, then Streamable HTTP — and the keys, limits and audit rows the traffic needs. - [RAG Architecture Diagram: the Boxes and What Each Arrow Costs](https://smartgate.network/industry/rag-architecture-diagram): A RAG architecture diagram in words: the index, retriever, reranker, context builder and loop controller, and what every arrow between them costs in tokens and latency. - [RAG vs Agentic AI: One Names a Pattern, the Other a System](https://smartgate.network/industry/rag-vs-agentic-ai): RAG names a retrieval pattern with no side effects; agentic AI names a runtime that plans, calls tools and acts. Where the line sits, and what each term is used for. - [RAG vs Agentic RAG: Choosing per Workload, Not in General](https://smartgate.network/industry/rag-vs-agentic-rag): RAG vs agentic RAG is a per-workload decision, not a preference: latency, cost, answer shape and the failure mode of each path, on one table you can decide from. - [What Is Agentic RAG? Retrieval as a Loop, Not a Pipeline](https://smartgate.network/industry/what-is-agentic-rag): Agentic RAG puts retrieval inside an agent's loop: the model searches, judges what came back, and searches again. What changes, what it costs, and when it is worth it. - [What Is MCP Model Context Protocol? Naming and Abbreviations](https://smartgate.network/industry/what-is-mcp-model-context-protocol): MCP and Model Context Protocol name the same thing, and the two spellings appear in different places on purpose: what the acronym promises and what it does not. - [What Is RAG Architecture? The Decisions You Cannot Defer](https://smartgate.network/industry/what-is-rag-architecture): RAG architecture is a set of design decisions: which components you can swap, how each swap trades latency against answer quality, and the failure every choice invites. - [What Is the Model Context Protocol? A First-Principles Guide](https://smartgate.network/industry/what-is-model-context-protocol): The Model Context Protocol is the wire format an AI host uses to reach tools it never shipped with. Here is the problem it solves, its three actors, one request end to end, and where it stops. ### By role - [FinOps Lead: Own the AI Bill — Measure, Alert, Report](https://smartgate.network/solutions/finops-lead): For the FinOps lead who owns the AI bill: a per-team quota percentage, 80% and 100% alerts, avoided-token dollars, and audit rows summed into a spend report you can sign off on. - [SmartGate MCP for Research and Decisions: A Personal Workflow](https://smartgate.network/solutions/smartgate-mcp-for-research-and-decisions): How an individual uses SmartGate MCP for research and decisions: one server entry in the assistant you already use, then search, fetch, deduplication, context and memory on every question.