# Gate > Gate is an accountability layer that sits in front of the AI models your agents call. Every request passes through one endpoint, where prompt injection is screened, credentials and personal data are redacted, spend is capped, tokens are compressed and cached, and the whole exchange lands on a tamper-evident record. Gate is built by Constellation Network. It works with the AI tools and subscriptions a team already runs: point an existing SDK or client at Gate and no other code changes. ## Product - [Gate](https://constellationgate.ai): What Gate is: screening, redaction, spend control, and a tamper-evident record on every request your AI agents make, without replacing the tools your team already uses. - [AI agent security](https://constellationgate.ai/security): How Gate screens requests for prompt injection and redacts credentials and personal data before they reach a model, with per-dataset benchmark results and the methodology behind them. - [Pricing](https://constellationgate.ai/pricing): Plans, and how model usage is billed. - [Gate Connect](https://constellationgate.ai/gate-connect-app): A menu-bar app that points the AI tools you already use at Gate. macOS, Windows, and Linux. - [Gate for OpenClaw](https://constellationgate.ai/openclaw-audit): An open-source OpenClaw plugin that records every tool call, message, skill, and cron an agent runs to a tamper-evident chain on your own machine. - [Blog index](https://constellationgate.ai/blog): Posts from the team building Gate. Every number ships with its methodology. ## Learn Plain-English explanations of how AI models, agents, and the systems around them work. Vendor-neutral reference, grouped by subject. - [Learn index](https://constellationgate.ai/learn): every explainer, grouped by subject. ### Requests and tokens - [What happens when an app calls an AI model](https://constellationgate.ai/learn/how-an-llm-api-request-works): A plain-English guide to how an LLM API request works: what an app sends, tokens, the context window, streaming, and what you pay for. ### Models - [Why models get things wrong](https://constellationgate.ai/learn/why-models-get-things-wrong): The distinct ways an AI model fails: invented facts, stale knowledge, drift in long documents, agreement bias, and run-to-run variation. What reduces the rate. - [How to choose a model for a task](https://constellationgate.ai/learn/how-to-choose-a-model-for-a-task): How to pick an AI model for a task: what differs between models, whether to start cheap or strong, and why retirement dates belong in the decision. - [Open-weight and closed models](https://constellationgate.ai/learn/open-weight-and-closed-models): What open weights means, why the licence matters more than the label, the three ways to run an open model, and what self-hosting requires. ### Cost - [What a request to an AI model costs](https://constellationgate.ai/learn/what-a-request-costs): How model pricing works: two separate rates, why a conversation costs more as it goes, what caching and batching change, and the line items that aren't tokens. - [How to reduce what you spend on AI models](https://constellationgate.ai/learn/how-to-reduce-llm-api-costs): A practical order of operations for cutting model spend: measure first, then the two changes that never touch your product, then the ones that trade something away. ### Agents and tools - [What an AI agent is](https://constellationgate.ai/learn/what-an-ai-agent-is): An AI agent is a model in a loop with tools, choosing its own steps. What separates it from a chatbot or a workflow, and where the cost and risk come from. - [How tool calling works](https://constellationgate.ai/learn/how-tool-calling-works): Tool calling lets a model ask your app to run an operation and use the result. The round trip, who executes what, and what the model knows about a tool. - [What MCP is](https://constellationgate.ai/learn/what-mcp-is): The Model Context Protocol standardises how apps expose tools and data to AI clients. What it defines, what it doesn't change, and what its spec says on security. ### Data and privacy - [What happens to the data you send a model](https://constellationgate.ai/learn/what-happens-to-your-data): Where your prompts go, who keeps them and for how long, whether they are used for training, and what differs between a consumer product and a business API account. ### Security - [What prompt injection is](https://constellationgate.ai/learn/what-prompt-injection-is): Prompt injection is text that reaches a model and changes what it does. Why it is structural, why tools make it expensive, and what reduces the damage. ### Infrastructure - [What an AI gateway is](https://constellationgate.ai/learn/what-is-an-ai-gateway): An AI gateway sits between your applications and the model providers they call. What the layer does, whose account pays for the tokens, and when it earns its place. ## Blog - [The Trade We Are Making](https://constellationgate.ai/blog/the-trade-we-are-making): The trade we make with AI matters more than how powerful it gets, and so does who answers for it. - [Building a product in the age of AI](https://constellationgate.ai/blog/open-poker-maintains-itself): How two AI agents help me keep a real side project moving after the first version ships. - [Delete the chat. The record survives.](https://constellationgate.ai/blog/delete-the-chat-the-record-survives): A CEO deleted his ChatGPT logs. - [Software is writing itself](https://constellationgate.ai/blog/software-is-writing-itself): A $40 dev board that became whatever I described. - [Frontier models fall to attacks you can script](https://constellationgate.ai/blog/frontier-models-fall-to-scripted-attacks): OpenAI's whole GPT-5.6 line complied with an injected send in every trial, Grok 4.6 fell on the morning of its release, and Claude Sonnet 5 fell in the deepest session we ran. - [AI Has Three Cost Centers. CFOs Need One Operating Model.](https://constellationgate.ai/blog/ai-has-three-cost-centers): Every AI workflow now carries three costs at once: the tokens and compute it burns, the human work it reshapes, and the governance needed to explain the result later. - [Your system prompt is not a security control](https://constellationgate.ai/blog/your-system-prompt-is-not-a-security-control): We wired a defended email agent to a test inbox and sent it one forged message. - [The Redeployment Problem](https://constellationgate.ai/blog/the-redeployment-problem): AI is replacing tasks rather than whole jobs, and the companies furthest along are redeploying people instead of cutting them. - [AI subscriptions will end](https://constellationgate.ai/blog/ai-subscriptions-will-end): Most of us met AI through a flat monthly subscription, and that price was never the real cost of the compute underneath it. - [Faster Isn't Efficient: Governing the AI Now Running Across Your Organization](https://constellationgate.ai/blog/faster-isnt-efficient): Every team is adding AI, and almost none are adding visibility. - [Pick your models. Own your rails.](https://constellationgate.ai/blog/pick-your-models-own-your-rails): Buy AI from a frontier lab and you get a whole operations department with the model: the screening you never see, the spend limits, the dashboard, the record. ## Reference - [Documentation](https://docs.constellationgate.ai): API reference, model catalogue, and integration guides. - [Sitemap](https://constellationgate.ai/sitemap-index.xml): every indexable URL on this site.