Every customer wants their numbers their way.
A governed analytics layer inside your product. Define metrics once; every customer filters, drills down, and saves their own views, and agents answer from the same context.
We build it on a sample of your data. You see it live in your product in a week.
“They can now more quickly prototype and test new use cases directly in our application with our clients.”
“The power of a full BI platform without locking us into rigid data models or charting libraries.”
“Our customers are thrilled with the data insights we’re able to provide.”
The charts were never the hard part.
You shipped a dashboard. Then every customer asked for it a different way: their sites, their contracts, their fiscal year, their export. Each request is small. Together they turn your CS and product teams into a report factory, and a frozen dashboard is out of date the day a customer signs a new contract.
AI can write the snapshot. It can’t run reporting for 200 customers who each want a different cut.
One definition. Every customer’s cut.
Every person, in every customer, gets an answer shaped to who they are: their sites, their role, their fiscal year, their saved views. Same number underneath.
Define each metric once.
On-time completion, margin per ton, yield. One governed definition in the semantic model, across every dataset and domain it touches. Every chart, export, question, and agent reads the same definition.
Every customer gets their own walled world.
Tenants are first-class: their own users, roles, groups, calendars, and number formats. Row, column, and connection-level security is applied on every query, so a customer only ever sees their own data.
They cut it their way. No ticket.
Inside your product, under your brand, your customers filter, drill, save their own views, ask questions in plain English, and schedule reports to their own teams. Your roadmap stays yours.
Any agent can write code. Only Semaphor governs the numbers.
Everything above is headless. Claude and Codex build dashboards and full data apps over API and MCP, against the same semantic model and the same tenant security your customers get. Faster builds, same governed answers.
Read the MCP docsA signed token and an embed URL.
Works in any framework, fully white-labeled, decoupled from your deploys. Go further with the Data App SDK when a customer needs more than a dashboard. Reads live from the database you already have. Cloud or self-hosted.
Live in a day. Still trusted a year on.
“Everything just works, which makes it easy to trust the product in day-to-day use.”
“Semaphor allowed our team to launch customer analytics in under a day.”
Priced for the platform, not per viewer.
Adding your 51st customer, or their 500th user, should not change your bill. Plans are flat by platform capability, cover embedded, internal BI, or both, and run on our cloud or yours.
Frequently asked questions
What is Semaphor?
Semaphor is a customer-facing analytics platform you embed in your software product. You define each metric once in a governed semantic model, and every customer filters, drills, and builds their own views of it inside your product, isolated per tenant. The same governed core also serves internal BI for your team and AI agents over API and MCP.
Can each of my customers see the numbers their own way?
Yes. Each customer is a first-class tenant with their own users, roles, calendar, and number formats. Inside your product they filter, drill down, save their own views, ask questions in plain English, and schedule reports to their own team, without a request to yours. Row-, column-, and connection-level security keeps every customer inside their own data.
Why not build customer reporting ourselves with AI?
AI makes the first dashboard cheap. The ongoing cost is everything around it: tenant isolation, permissions, consistent metric definitions, exports, scheduled delivery, and a new request from a different customer every week. Semaphor is that layer, and coding agents build on top of it through MCP, so you keep the speed without owning the maintenance.
What is agent-native BI?
Agent-native BI is business intelligence architected so AI agents are first-class users: they query a governed semantic layer over API and protocols like MCP, and every answer resolves against approved definitions and is verified against a live query rather than guessed. Semaphor is built on this architecture from the ground up.
How do coding agents build dashboards in Semaphor?
Semaphor ships an MCP server that agents like Codex and Claude use to discover your data, propose views, and build dashboards and data apps. Every metric is grounded in your semantic model and verified against a live query before it ships.
Can I embed Semaphor dashboards in my product?
Yes. Embed with a zero-dependency iframe or go native with the React, Vue, and Web Component SDKs. Embedded dashboards are white-label and multi-tenant, with row-level security applied per tenant on every query.
Does Semaphor replace my internal BI tool?
Yes. Teams use Semaphor for governed data exploration, dashboards, scheduled briefings, and plain-English questions, without per-viewer seat pricing.
Which data sources does Semaphor support?
PostgreSQL, MySQL, BigQuery, Snowflake, ClickHouse, DuckDB, SQL Server, Amazon S3, and REST APIs. Queries run live against your source, with no ETL pipelines and no data duplication.
Can Semaphor be self-hosted?
Yes. Semaphor runs on your own infrastructure with Docker, keeping data inside your network for GDPR and HIPAA requirements.
Show us the report your customers keep asking for.
We build it on a sample of your data and hand you the embed. One week. If it doesn’t close tickets, you’ve lost nothing.
Or explore the live demo