Outreach without the growth-hack playbook.

Cold outreach has a bad reputation because most tools optimize for volume over judgment. These four principles are the constraints that keep FirstContact from becoming another blast tool.

01 · Evidence before inference

Investor mandates, portfolio patterns, and contact details remain linked to their public source and discovery date. A claim without a source is treated as unverified, not as fact. This is what lets a founder trust a match reason instead of taking it on faith.

02 · Permission at the edge

No hidden auto-send. Live email requires operator enablement, policy checks, suppression screening, and a specific human approval of the exact recipient and exact message — every time, not just the first time. See Responsible outreach for the full gate list.

03 · Founders own the narrative

AI can structure and draft, but it cannot invent traction, impersonate a founder, or overwrite their context. The founder's own words about their traction, impact, and local context stay theirs — a model's job is to organize them, not replace them.

04 · Open by default

Portable data, inspectable scoring, replaceable providers, and documentation deep enough to run independently. Every provider — Exa, OpenAI, Resend — is an adapter behind a boundary, not a dependency baked into the product's identity.

These principles are enforced in code, not just stated: see the system for how each layer fails closed, or read how it works end to end.