Alli MCP Server is a Model Context Protocol server that connects AI assistants and agents to the Alli operating system. It turns Alli’s APIs into callable tools, so agents can query governed data, build dashboards, manage datasources, and trigger workflows from plain-English requests, with OAuth 2.1 security and no custom integration code.
Teams want AI assistants that answer real questions and take the next step. Without governed access to the data and services behind the work, every new use case turns into an integration project.
Every new AI use case adds another client, wrapper, and auth flow to build, then maintain when the API changes.
Reviewing a client’s full setup across datasources, dashboards, and workflows means clicking through the interface one page at a time.
Agents connected through one-off scripts may see more than they should, and nobody can trace what they touched.
Turn the Alli API into tools any MCP-ready agent can call, on your governed data, with OAuth handled for you.
Add the Alli MCP server to an agent, and it pulls from the same governed data as every Alli report, with the same metrics and definitions.
For your teamAnswers from an agent match the numbers in the dashboard.
Explore Semantic LayerAlli services show up in the agent as tools you can turn on. Tool definitions update as Alli’s API specifications change, so there’s no wrapper to maintain.
For your teamDevelopers stop maintaining one-off integrations for each AI use case.
An agent can call several Alli tools in a row, from refreshing a creative to pulling ad details to generating images and suggestions, to finish a cross-team task.
For your teamA request that used to take several handoffs gets done in one conversation.
Explore Generative WorkflowsAgents query governed data, so they use the same metrics and definitions as your reports.
Agents can see how clients, datasources, dashboards, and workflows relate to each other, so one request covers the whole setup.
With callable tools, agents can launch a workflow or build a dashboard in the same conversation where the question was asked.
Agent-ready means permissioned, so agents only reach the tools and data the signed-in user is allowed to use.
Who uses it
Ask for a pacing snapshot.
Get spend pacing across active datasources for a client in one request, with no exports to pull.
Catch schema drift early.
Compare a datasource’s current schema to its expected definition before a report breaks.
Prototype an API call conversationally.
Try a call in plain English before committing production code, then build on the same tools.
See a client’s whole footprint.
Review datasources, dashboards, and workflows before an audit or planning session, on a governed path for AI adoption.
Every request carries the signed-in user’s own authorization, and teams decide which tools each agent can see.
Add the server, choose the tools, and keep the count where the agent works best. OAuth 2.1 with PKCE and token forwarding run automatically, so agents act within each user’s access.
For your teamSecurity teams get agent activity inside the governance they already run.
Refresh a single creative by its ID, pull the details of a specific ad, generate new images and suggestions, or see which Alli servers and tools the agent has turned on.
For your teamTeams build trust with read-only questions, then move to actions.
Connect an MCP-compatible assistant, ask in plain English, and the agent calls the Alli tools your permissions allow.
Add Alli MCP Server to an MCP-compatible assistant or agent and sign in with OAuth.
The server lists the Alli tools your permissions allow, with what each one does.
For example, “Show me spend pacing across active datasources for this client.”
It chooses the right tools, chains them if needed, and sends each request with your token.
Read the result, then launch a workflow or build a dashboard in the same conversation.
Alli MCP Server covers more than 20 Alli services. These are the ones teams reach for first.
Need a service that isn’t listed? Tools follow Alli’s API specifications as they change. Ask about a service
Hand-built API integrations need code for every use case, and working screen by screen doesn’t scale to big questions. Alli MCP Server turns the API into governed tools agents can call.
| Manual API integration | UI-only work | Alli MCP Server | |
|---|---|---|---|
| Setup | Write clients, wrappers, and auth for each use case | None, but every task is manual | API specifications become callable tools automatically |
| Keeping current | Code breaks when APIs change | Not applicable | Tools update as API specifications change |
| Multi-step tasks | Depends on the script | Slow, screen by screen | One plain-English request across services |
| Who can use it | Developers | Anyone, one task at a time | Practitioners and developers through AI assistants |
| Security | Varies by implementation | Platform login | OAuth 2.1 with PKCE, token forwarding, stateless requests |
| Data foundation | Whatever the script queries | What’s on the screen | Governed Alli data with business context |
Alli MCP Server keeps every agent request inside your existing governance, using modern authorization and no stored tokens.
Security: the Alli platform holds a current SOC 2 Type II report and is aligned with the NIST Cybersecurity Framework (CSF). Request the report See security details
Modern authorization for every client connection.
Each request carries the user’s own authorization, so agents act within that user’s access.
Every call is a fresh transaction with no tokens stored between requests, and tool definitions stay in step with Alli’s API specifications.
The ideas behind Alli MCP Server
What MCP is, how Alli MCP Server stays secure, and how teams use it.
Alli MCP Server is a Model Context Protocol server for the Alli operating system. It turns Alli’s APIs into callable tools so AI assistants and agents can query governed data, build dashboards, manage datasources, and trigger workflows from plain-English requests. It covers more than 20 Alli services and secures access with OAuth 2.1.
Model Context Protocol, or MCP, is an open standard for connecting AI models to external tools and data. A system exposes its capabilities through an MCP server, and any MCP-compatible AI client can discover and call them. It gives agents a shared way to use software, so teams don’t build a one-off integration for each tool.
Alli MCP Server works with AI clients that support Model Context Protocol, including AI coding assistants and agent frameworks. Because MCP is an open standard, teams can use the assistant they prefer. Each one sees the same set of Alli tools, governed by the same permissions.
Alli MCP Server uses OAuth 2.1 with PKCE for authorization and forwards each user’s token with every request. Requests are stateless, so tokens aren’t stored between calls. Agents can only reach the tools and data that the signed-in user is permitted to access, which keeps agent activity inside your existing governance.
Both, depending on permissions. Agents can read data and inspect configurations, and they can also take actions like triggering workflows, building dashboards, or managing datasources when the user has rights to do so. Teams decide which tools to enable, so read-only setups are possible where that’s preferred.
The Alli API needs code: an HTTP client, authentication, and maintenance as endpoints change. Alli MCP Server wraps that API as agent-ready tools, handles authentication, and updates the tools automatically when API specifications change. Developers still have the API when they need it, and most AI-assisted tasks no longer need custom integration.
Ask Alli is the conversational interface built into Alli. Alli MCP Server brings Alli’s capabilities to the AI assistants and agents your teams already use outside Alli. Both work from the same governed data foundation, so the answers stay consistent whichever interface your team chooses.
Start with read-only questions that take too many screens to answer today, like a client’s full platform footprint or a pacing snapshot across datasources. Once the team trusts the answers, move to actions such as drafting dashboards and triggering workflows, and use it for prototyping new integrations.
Book an intro and we’ll connect an AI assistant to Alli live, then show you what it can find and do with one plain-English request.