The framework you build with. The runtime that runs it.
SEAM Framework, Standard and Scale
Framework
Open source. Apache-2.0.
Free
Individual and internal use
Author SEAM definitions in YAML, validate, compile and resolve them from your terminal. Free for individual practitioners and internal organisational use.
- check Local CLI on npm
- check Author, validate, compile, resolve
- check Apache-2.0 licensed
# Install
npm install -g @measurelab/seam-cliStandard
Cloud hosted. Implemented by Measurelab.
£900/mo
Or £750/mo billed annually (£8,970/yr, save £1,830)
The managed runtime in front of your agents. Workshop-led implementation, every call resolved through your definitions, complete audit trail.
- check Up to 7 connectors from our library or your own MCPs
- check Unlimited team members
- check Canvas and the Console included
- check Per-user OAuth with git-worktree workspaces
- check Audit Log to BigQuery + AI Act trail
- check Workshop-led implementation
Scale
Bespoke connectors. Custom build-out.
Custom
Quoted against your sources
For organisations with more than seven sources, or who need connectors we haven't built yet. Most of the build effort goes into the new connectors and the intelligence model on top.
- check Everything in Standard
- check More than 7 connectors
- check Custom connector development for sources we haven't built
- check Intelligence-model build-out on top of new sources
- check Dedicated implementation lead
What's included
| Framework | Standard | Scale | |
|---|---|---|---|
| Author definitions in YAML | check | check | check |
| Validate, compile, resolve locally | check | check | check |
| Visual map and editor (Canvas in the Console) | - | check | check |
| Production MCP server | - | check | check |
| Per-user OAuth with git-worktree workspaces | - | check | check |
| Source connectors | - | Up to 7 | 7+ |
| Custom connector development | - | - | check |
| Downstream MCP proxying for any custom source | - | check | check |
| Audit Log with governed / ungoverned / resource-gap verdicts | - | check | check |
| Audit Log to BigQuery (full history) | - | check | check |
| AI Act audit trail | - | check | check |
| Implementation engagement | - | check | check |
| Team members | Single user | Unlimited | Unlimited |
| Get started | npm install | Book a demo | Book a demo |
Book a demo. Chat about where you fit.
Thirty minutes with someone from our team. We'll show you SEAM Canvas in action, walk you through the benchmark - 95% correct vs 44% on raw MCPs, 2.3× cheaper, 4× more accurate on governed metrics - and ask about your organisation: your AI agents, your data sources, the governance gaps. You'll see what an audit trail by default looks like in practice. No commitment. We tell you honestly whether SEAM fits.
Or email hello@measurelab.co.uk
Thanks. We'll be in touch within one working day.
Compliance regimes for AI agents are arriving fast.
The EU AI Act, ISO 42001, GDPR's AI extensions - all of them require what SEAM produces by default:
- check_circle An inventory of agent actions
- check_circle Traceable data flows
- check_circle Behavioural drift detection
- check_circle A complete audit trail
Building this yourself is a multi-quarter project. SEAM ships it on day one of an engagement.
Common questions, honest answers.
What counts as a connector? expand_more
A connector is one source SEAM resolves against - for example BigQuery, GA4, HubSpot, Slack, Notion or a custom MCP server we proxy. Standard includes up to seven from our pre-built library or your own systems wired through MCP. Connector count is the main difference between Standard and Scale because building new connectors we haven't shipped before is where most of our AI-augmented implementation time goes.
What if we need more than seven connectors? expand_more
That's Scale. Pricing is bespoke because it depends on how many new connectors we'd build for you and how much intelligence-model work goes on top. Book a demo and we'll quote the engagement against your sources.
Where do I get the CLI? expand_more
npm install -g @measurelab/seam-cli. Apache-2.0. Validate, compile and resolve definitions locally. The runtime that runs production SEAM in front of agents stays Measurelab-managed - talk to us if you need that.
What does an engagement involve? expand_more
It starts with a 30-minute demo and chat. We show you Canvas, walk through the benchmark numbers and what they mean, and ask about your organisation - your AI agents, your data sources, the governance gaps. You see the audit trail in action. From there we shape the right engagement together.
Can I self-host? expand_more
The CLI runs locally - you can author, validate and resolve against your own model on your machine. The production runtime (multi-tenant OAuth, downstream MCP proxy, audit pipeline) is Measurelab-managed only. We're not currently distributing the runtime for self-hosting.
How does SEAM compare to Atlan, Collibra or dbt's MCP? expand_more
Data catalogues are for humans browsing. SEAM is the intelligence layer for agents resolving. Use both. dbt MCP exposes your dbt project to agents; SEAM is a layer above, governing how those calls resolve regardless of source.
Are you taking partners? expand_more
Yes. The Get involved page covers three lanes: the founding customer cohort, design partners and a waitlist for consultancy or agency partners delivering SEAM engagements through accreditation.
What's the founding-customer arrangement? expand_more
The first cohort joins under preferential terms - direct line to the SEAM team, priority on roadmap input, named in our public materials, regular feedback sessions. Full details on the Get involved page.