Contact

Different questions.
Different doors.

Enterprise evaluation, partner inquiry, press and analyst briefings, and technical integration questions are different conversations with different people at grāmatr. Choose the one that matches where you are.

grāmatr is the intelligence layer: it classifies the request, delivers the right context so the model returns better answers for fewer tokens, enforces policy, attributes every token, and compounds what it learns — before every model response.

Enterprise Evaluation

The first conversation is an architecture review — not a demo.

We want to understand your current AI deployment, where the context, attribution, and policy gaps live, and whether grāmatr belongs in your request path. Forty-five minutes. No deck, no product tour, no pitch. We map your stack and tell you exactly where the intelligence layer is missing — or confirm it is not.

Step 1 of 2Four fields. About fifteen seconds. The rest is optional.

A sentence is plenty. It tells us who to put on the call and what to look at before it.
What happens after you submit

An infrastructure engineer at grāmatr responds within one business day to confirm the architecture review. The session is 45 minutes. We map your current deployment, identify where the intelligence gaps live — classification, context, attribution, policy, and compounding — and tell you exactly where grāmatr belongs in the request path — or doesn't. No deck. No follow-up sequence unless you ask for one.

This form routes directly to the grāmatr engineering team. We do not use a sales development sequence. A person reads this and responds.

Partner Inquiry

You need infrastructure you can stand behind with clients.
We need to understand what you're delivering.

We are building three partner tracks: reseller, embed and OEM, and managed deployment. Every track starts with the same question — what are you trying to deliver to clients, and what do you need to defend it when a client's procurement team or regulator asks? That is a 30-minute conversation. Anneal — the first application built on grāmatr — means you do not have to build anything to start: deploy the full grāmatr intelligence layer to your own consultants on day one, then accelerate every client engagement on the same infrastructure.

What happens after you submit

Our partner team responds within one business day. If you selected the scope call, we will send three time options for a 30-minute conversation — we want to understand your client portfolio and where grāmatr fits your delivery model before we discuss terms. If you selected the partner overview first, we send it with a note about next steps. No automatic CRM sequence — a person reads this.

Partner conversations begin with a scope call, not a term sheet. We want to understand your practice before we talk about economics.

Press & Analyst

We do briefings. We do not have a PR agency.

On-record comment requests, background briefings, analyst briefings with NDA, and interview requests all route here. You are reaching the team that built the system — not a communications intermediary. On-record comment requests receive a response within 24 hours. We do not embargo product announcements without advance written agreement.

What happens after you submit

On-record comment requests receive a response within 24 hours. Briefing and interview requests are scheduled within two business days of your stated deadline. Analyst briefings that require non-public product detail are conducted under NDA — note that in your request and we will send an NDA for countersignature before the briefing is scheduled.

If your deadline is under 24 hours, note it clearly in the coverage field. We will do our best — we cannot guarantee response time for same-day requests that arrive without advance notice.

Technical Question

API architecture, deployment, SDK, integration design — the engineering team reads this.

Technical questions about the grāmatr API, deployment configuration, SDK integration, and request-path architecture are handled directly by the engineering team. If your question is better answered in a 30-minute call than by email, we will say so and schedule one. If it belongs in the documentation, we will tell you where.

Include error messages, request shapes, or configuration fragments if they are relevant. Do not include secrets or credentials in this field.
What happens after you submit

Technical questions are reviewed by the grāmatr engineering team and responded to within two business days. If your question is better answered in a call, we will propose a 30-minute time. Complex integration questions may require a technical briefing session rather than email — we will tell you that rather than sending a reply that doesn't answer the question.

This goes directly to the engineering team — not to support ticketing. Specific questions get specific answers.

930,000+
memories assembled and delivered in production. Real applications run on grāmatr — see the production record.
<100ms
pre-classification latency. The intelligence layer does not slow the AI estate.
99.9%
SLA. Fail-open design — if grāmatr goes down, requests flow through. No single point of failure added to your infrastructure.
SOC 2
Type I targeting October 15, 2026. Type II targeting April 15, 2027. Row-level security enforced at the database layer.