Introducing Headless Amplitude. Say bye bye to the user interface. Agents are now first class users of Amplitude. - All core areas of Amplitude are now accessible via MCP server and we've optimized the interface for minimal token consumption - Developer CLI to manage events, experiments, flags, and projects... so you can make changes in Amplitude deterministically - Service accounts so bots and scripts can query Amplitude on their own accounts In August alone, 6.3M MCP tool calls ran through Amplitude, 5x what we saw in March. Querying product data is now our fastest growing agent use case. You shouldn't have to log into the 100th SaaS dashboard to understand what your users are doing. Amplitude can show up in Codex, Claude, Cursor, or whatever agent your team is already using. We’re building toward that future with MCP, developer CLI, service accounts, and agent readable experiences. Use Amplitude without using the UI.show more

Congrats on @eric_nakji @tommydotkeeley @Jacob__Newman @ChanakaP30674 on the launch of Headless Amplitude! Details here:

Making @Amplitude_HQ easier and better to use for agents!

@samseely @cjbell_ checkout the demo vid, very aligned with our chats last night at the @knocklabs meetup!

confirmed, using it daily, works great so far. codex gets the data and also quickly adds missing events/properties in our code

bye bye user interface is bold but honestly fits where software is heading

Congratulations to the team for continued to iterate on my new favorite way to use Amplitude!

Product data in more places means more power and more efficiency! Super excited for this.

As someone that works in the security space, I'm thrilled that we can now separate the identities of agents from humans! I think Enterprise customers will be thrilled too.

Amplitude finally making analytics work natively inside agentic workflows

Can you guys increase session lifetime for the MCP? I love the MCP, but it’s a pain having to re-auth every couple of days

I never have to use a UI again!

Just yesterday chatted with a user who shared how easy it was for her to automate monitoring onboarding funnels with Amplitude MCP combined with Datadog, Snowflake and Slack MCPs on Claude, where reports would go directly to her team on Slack

Yall are cookin

do agents get seats too

the interesting half is not the mcp server, it is who owns the definitions once an agent is the one asking. we run an analyst agent and the biggest single quality jump came from making it refuse a question when the join was ambiguous instead of returning a confident number. a ui at least made a human pause at that moment. does headless return an answer when the metric is defined two ways in the warehouse?

😬

Buy the stock!

Love it. More room for growth as well once you’ve started to optimise you mcp for organic discovery

the bottleneck in analytics was always translating a question into the query builder, so this fits well. what you lose is wandering. a dashboard showed you the question you should have asked.

6.3M MCP calls. does a typical funnel query return a small schema, or does the agent still pull the event dictionary first?

The MCP layer is the interesting part here. Agents can act on analytics directly.

As usage moves into agents, are you seeing a different person champion Amplitude inside the customer? I’m curious whether the buying conversation is changing too, or whether the same buyer is simply getting value through a different workflow.

The key shift is making product analytics an agent-readable control plane: MCP for querying, CLI for deterministic changes, and service accounts for unattended operation.

Agent native analytics could eliminate a lot of dashboard hopping