About
About Red Orca
Red Orca is a European software company that builds AI agents. Its three products share one account: Dolphin, an AI voice and email agent for outbound sales; Shark, organizational memory that tracks what a team decided and who owes what; and Oyster, a cybersecurity AI assistant.
What we are building
AI agents that finish the job — and show their work.
Red Orca builds specialized agents that work inside the systems a business already runs: its phone lines and mailboxes, its documents, and its security work. Each agent has one job. Each one should be able to show what it saw, what it did and why — because an AI system a business can't inspect is one it can't safely rely on.
Products
Three agents. One account.
Dolphin, Shark and Oyster are separate agents with separate jobs. They share one account and one sign-in.
AI Voice and Email agent
Dolphin
Dolphin is Red Orca's AI voice and email agent for outbound sales: it calls your leads within the hours you set, holds a two-way conversation to qualify them, keeps a full transcript of every call, and runs cold email campaigns from your own mailbox.
Organizational Memory
Shark
Shark is Red Orca's organizational-memory software: it reads the emails, notes and documents a team already produces and keeps track of what was decided, who owes what and where documents contradict each other, citing the source sentence for every answer.
Cybersecurity AI assistant
Oyster
Oyster is Red Orca's cybersecurity AI assistant: a chat for security engineering questions — how attacks work, how to detect and prevent them — plus, with a free account, questions over your own files with cited answers, red-team planning limited to a scope you declare, and a SOC Copilot workbench for the evidence you bring.
Private by design
Data exposure should be a conscious architecture decision, not a hidden default.
Approval before consequence
Drafting and recommending are different from sending, publishing or changing records. Consequential actions need explicit policy and, where appropriate, a person's approval.
Evidence over confidence
A polished answer is not proof. Systems that work with company knowledge should show the source behind an answer and admit when the evidence is insufficient.
Least privilege
An agent should receive only the tools, data and permissions the task needs.
Traceability by default
A business should be able to see what the system saw, what it proposed, what was approved and what happened.
In practice
What that means in the products today.
- Dolphin always confirms it is an AI when asked and never claims to be human, and a campaign cannot start until you confirm the people on it agreed to be contacted.
- Shark cites the sentence behind every item it extracts, and nothing it extracts counts as fact until a person confirms it.
- Oyster explains and plans. It has no access to the systems you ask about, and its models run on infrastructure we operate ourselves, in the EU.
- We say what is live, what is being tested and what is still planned — and correct our own pages when they fall behind the product.
Red Orca is built in Europe.