About ClariLayer
ClariLayer is the context layer that checks itself — reconciled against your source, not just asserted. Built for the individual analyst, delivered over MCP. Connect it to your AI coding agent and it stops making the same data mistakes — wrong table, wrong join, refunds counted in revenue — session after session. It comes from the day-to-day reality of working against your own messy warehouse or CRM semantics with an agent that forgets.
Analysts have moved into AI coding agents — Claude Code, Cursor, Codex — and they are productive there. But the agent has no durable memory of your data. Every session starts from scratch: which table is the real one, how two tables actually join, why this status code means a refund, that one customer who breaks every rule. You explain it, you get an answer, and tomorrow you explain it again.
The common workaround — a hand-written CLAUDE.md of definitions — helps, but it has the same trust problem as the original numbers: it is just asserted text. Nothing checks that what you wrote down still matches what the warehouse returns or how a live CRM is configured and used. So the agent confidently repeats a definition that quietly drifted, and you only find out when two reports disagree.
That disagreement is the acute moment, and it is the research-validated #1 analyst pain: why don't these two numbers match? Today the honest answer is a slow manual audit. ClariLayer was built to make that moment fast — to reconcile one saved definition against supported live evidence and flag a caveat when the declaration and the evidence disagree.
Operator Credibility
ClariLayer grounds credibility in the work itself: the analyst's daily grind of re-explaining data to an agent, the trust break when the numbers disagree, and the honest path from one person's context to a team's. The proof is the workflow and the evidence trail, not borrowed brand marks.
You live inside Claude Code, Cursor, or Codex against your own warehouse or operational systems — often messy, ungoverned, stitched across sources. Every new session you re-explain the same things: which table is real, which join is right, what a CRM stage means, that refunds don't belong in revenue. The agent forgets; you repeat yourself.
A hand-typed CLAUDE.md of claimed definitions has the same problem as the original numbers — it is just asserted text. The acute moment is the one analysts feel most: two reports, two answers, and no fast way to ask "why don't these two numbers match?" That is the pain ClariLayer was built around.
The same engine that grounds one analyst's context becomes shared, owned, governed team context — ownership, approval, the one right metric. That team layer exists today: the Governed Context Edge is built and in private pilot, onboarding design-partner teams now.
ClariLayer is a personal context layer for your AI, delivered over MCP. Connect claude.ai over OAuth, or install it into Claude Code, Cursor, or Codex, and your agent can recall the right context in-flow, mid-task. There is no destination app to visit and no team account to provision — single-player by design: one analyst, your own data, your own agent.
It runs on four verbs. Bootstrap populates your context from five shipped source kinds: SQL, dbt, CLAUDE.md, a dictionary/codebook, or a semantic model. Remember saves a new definition, join path, or gotcha so it survives across sessions. Recall pulls the most relevant context for the task at hand. And reconcile checks one compatible saved definition per explicit call. Warehouse definitions accept actual_sample evidence from agent-run SQL, including optional preview rows. HubSpot definitions accept bounded crm_evidence property metadata and aggregate distributions, with CRM rows forbidden. HubSpot reconcile is generally available. In this personal MCP path, we never hold source credentials, execute SQL, or call HubSpot ourselves.
We are deliberate about what we claim. Today reconcile records a caveat on a mismatch or leaves an entry asserted — it does not stamp a definition “verified.” The stronger status remains gated off, with no public release timeline. The present-tense promise is reconciled, evidence-backed, caveat-aware context — checked against supported evidence, not blindly asserted.
This is where we start, not where it ends. The same engine that grounds one analyst's context is the bridge to the team: the Governed Context Edge promotes the definitions you reconciled into shared, owned, governed team context — ownership, approval, the one right metric. That edge is built and in a hand-run private pilot today, onboarding a few design-partner teams.
Your agent shouldn't start from zero or a cold, hand-typed CLAUDE.md. Bootstrap the work you already have from five shipped source kinds: SQL, dbt, CLAUDE.md, a dictionary/codebook, or a semantic model. Day-1 value, not an empty store.
Each explicit reconcile grounds one compatible saved definition against agent-supplied evidence. Warehouse definitions use actual_sample, which may include optional preview rows; HubSpot definitions use bounded, recursively row-free crm_evidence. In the personal MCP path, we hold no source credentials, execute no SQL, and make no HubSpot call. A mismatch becomes a caveat; the entry otherwise stays asserted.
ClariLayer lives inside the agent you already use. One command installs it as an MCP server into Claude Code, Cursor, or Codex; from then on your agent has an in-flow recall tool it can call to pull the right context mid-task. No destination app to visit — it rides along.
Every correction, every reconcile, every note quietly persists. Your agent grounds on more of your context over time and gets progressively more right about your data. The longer you use it, the more it would sting to lose — the context you build is the moat.
Connect ClariLayer to claude.ai, Claude Code, Cursor, or Codex and give your agent durable, reconciled context about your data. Or bring the team: the Governed Context Edge is in private pilot — request early access.
Questions? Email us at support@clarilayer.com.
We use privacy-friendly analytics
With your consent we use PostHog and Vercel Analytics to understand how ClariLayer is used so we can improve it. We never sell your data. Errors are always monitored (without analytics) so we can keep the app reliable. You can change your mind anytime.