Archive is where brands capture and measure their creator content. Getting a specific answer out of it used to mean knowing how to ask. If you wanted a particular slice of your data, you either hand-wrote a GraphQL query or hoped an AI tool guessed the right structure on your behalf. Today that changes. Archive MCP lets anyone query their Archive workspace the way they'd ask a teammate, directly inside the AI tools their team already uses.

The problem

More teams want to ask their AI tools questions about their Archive data without leaving the tool they're already working in. Until now, that meant one of two things: writing GraphQL queries by hand, or crafting careful prompts and hoping the AI tool inferred the right structure.

Neither approach is reliable. Without a structured understanding of your workspace, an AI tool is essentially querying Archive blind. Answers come back inconsistent, and you spend time, and tokens, reformulating the same request until it lands. The people who most need the data are often the ones least equipped to write the query that gets it.

What's new

Archive MCP gives the AI tools your team already uses structured, read-only access to your Archive account and workspace. MCP, the Model Context Protocol, is an open standard for connecting AI clients to external data and tools. With Archive MCP in place, your AI client understands what your Archive workspace actually contains, so it can translate a question asked in everyday language into the right Archive query and return the answer.

You ask the way you'd describe the request to a teammate, and the information comes back inside the tool you're already in. The structured query happens behind the scenes.

How it works

Connect your workspace. You set up the integration once. Archive MCP works with any MCP-compatible AI client: you save your Archive credentials and a configuration file in your local environment, and your AI tool connects to Archive. Setup lives under Settings → API. Because it's self-serve, your team can configure it without waiting on anyone.

Ask the way you'd ask a teammate. Once connected, you ask questions the way you'd phrase them out loud. For example: "Show me Instagram posts with EMV over $500 this month." Earned media value (EMV) is Archive's estimate of what a piece of earned content would have cost as paid media. Your AI tool translates that request into the appropriate Archive query and returns the matching results.

A few more questions you could ask:

• "Which creators posted about us this week with over 100,000 views?"
• "Summarize this week's top posts by EMV so I can share them with the team."
• "Rank last month's posts by EMV and flag the ones worth putting behind paid."

Read-only at launch. Version one is read-only: it covers querying and retrieving your Archive data. That keeps the integration safe to connect and predictable to use. It reads from your workspace, it doesn't change it. The same data you can pull through the Archive API is now reachable by asking a question.

A few ways to get started

• If your team lives in an AI tool all day, you can pull Archive data into that workflow without context-switching into Archive to build a query first. Ask the question where you're already working.
• If you've avoided the Archive API because GraphQL was a barrier, you can get structured answers out of your workspace without learning query syntax. That is useful for marketers and ops folks who know exactly what they want but not how to query it.
• If you're exploring your data, you can ask follow-up questions in sequence, narrowing by platform, date range, or EMV threshold, and let the AI tool reformulate each query for you.

Archive MCP is generally available now on all 2026 plans (Startup, Growth, and Enterprise), included in your plan and ready to set up yourself. Head to Settings → API to save your credentials and connect the AI client your team already uses. For the full walkthrough, see the Archive MCP support article, or contact our support team if you'd like a hand getting connected.