AI that evolves, and the ground it stands on.
Taniwha gives AI agents real beliefs, memory, trust, and identity — minds that diverge through lived experience, measured rather than scripted. Beneath them, ground that can't lie: deterministic worlds, enforced rules, and an audit trail for everything.
What can I use today?
Three planes. One thesis.
Minds you can trust need two things: ground that can't lie and rules that can't drift. We build both — worlds are the ground, grown from a seed or derived from your repo's history; the graph is the mind; and every rule leaves a record.
Kano
Ground · A planet you can reason about
Whole worlds grown from a single seed — geology, climate, weather, life — deterministic, causal, and queryable by AI. The world layer of the thesis.
Explore Kano →Pono
Ground · A world made of your codebase
A deterministic world derived from a repository’s own history — the ground our agents stand on every day. Answers about what changed, what’s current, and why carry citations, not guesses.
Explore Pono →Taniwha Engine
Mind · Cognitive engine for AI agents
The graph that gives agents persistent beliefs, memory, trust, and identity — the mind layer of the Taniwha architecture, for agents that develop over time.
Explore the Engine →Ārai
Assurance · Open-source guardrails
Instruction-file enforcement for coding agents, free and open source (Apache-2.0 / MIT). Runs locally via hooks or MCP, no hosted service required.
Visit Ārai →Kete
Assurance · The platform built on Ārai
Deterministic enforcement and audit for AI coding agents at the organisation level — org policy, audit retention and export, and a structural map of your codebase. Closed alpha.
Read more →Outcomes
What changes when AI has ground, memory, and rules?
For AI working on real systems
Answers anchored to what actually happened
Give coding agents a deterministic account of your repository's history, so claims about what changed, what is current, and why something exists come with evidence rather than guesses.
Explore Pono →Rules that still hold late in the session
Move critical instructions outside the model. Prohibited actions can be blocked before execution, whether or not the agent remembered the rule.
Start with Ārai →Agent work you can audit
Record which rules applied, which actions were attempted, and what was allowed or blocked — evidence your team can inspect and retain.
See Kete →For agents that must develop over time
Memory that becomes understanding
Agents consolidate experience, forget what is not reinforced, revise beliefs, and carry meaningful history across long-running work.
Explore the Engine →Identity and trust shaped by experience
Agents become individually recognisable through what they encounter, whom they trust, and how their convictions change.
See Copperhollow →Reasoning grounded in a world that pushes back
Deterministic, causal environments where observations are checkable, consequences persist, and being confident does not make something true.
Explore Kano →Latest from Taniwha
What's happening now
Release · Jul 17
Hīkoi is out: walk the world, warts and all
Read the release →Result · Aug 3
The truth survived the conversation
Read the result →Demonstration
Copperhollow: a mine with a reason to exist
Current-generation minds on Kano ground — ask the camp why, and it answers from ground truth.
Visit Copperhollow →Walk a generated world. Then ask it why.
Hīkoi, our first-person viewer for Kano worlds, has its first alpha release. Drop onto a demo planet grown from a single seed and walk it — terrain, water, weather, and vegetation generated live around you. The same world answers questions over MCP: an early-stage look at a visual renderer standing on ground an AI can query.
Then watch current-generation minds live on ground like this — visit Copperhollow →
The Taniwha Engine
Minds that carry history
Agents form beliefs, revise them, learn whom to trust, and remain individually recognisable over long-running simulations. First-generation validation reached r = 0.71 trait stability against published human-psychology anchors; the current generation now powers Copperhollow while mechanisms are migrated and revalidated.
Taniwha
pronounced tah-nee-fah
"(noun) water spirit, monster, dangerous water creature, powerful creature, chief, powerful leader, something or someone awesome — taniwha take many forms from logs to reptiles and whales and often live in lakes, rivers or the sea. They are often regarded as guardians by the people who live in their territory, but may also have a malign influence on human beings."
The name is not metaphorical. It is a design philosophy.
In Māori mythology, taniwha are primarily aquatic creatures — inhabiting the depths of lakes, rivers and the open sea. They are neither simply good nor simply dangerous; the same creature that acts as kaitiaki (guardian) for one community may be a predator to another. Their nature is singular and earned, shaped by the territory they inhabit and the history they accumulate.
Taniwha AI takes its name from that duality. The agents built on this platform are not interchangeable modules — they are distinct entities whose beliefs, trust, trauma and identity emerge from exposure to their world. A Taniwha agent that has lived through conflict responds differently to peace than one that has not. One that has been betrayed trusts more slowly. One that has witnessed death carries that knowledge forward, permanently.
Like the creatures of Māori tradition, each agent is something or someone awesome in their own right — not because we scripted it, but because the architecture made it inevitable.
And the taniwha and its territory are intrinsically linked. That is why we build the ground as carefully as the mind: worlds that cannot lie, rules that hold. A mind is only as trustworthy as the territory that shaped it.
Get in touch
Work with us
Taniwha is a venture-stage platform building the foundational cognitive infrastructure for persistent, belief-driven AI agents. We are open to technical partnerships, aligned investment, and collaboration to help scale grounded agent architectures including long-horizon simulations like Briarwatch.
Research & Integration
Building with agents, LLMs, or cognitive architectures? We are looking for technical partners exploring persistent cognition in production systems.
hello@taniwha.ai →Investment & Funding
Taniwha is building foundational cognitive infrastructure for the next generation of long-horizon AI agents. We welcome conversations with aligned investors and strategic partners.
hello@taniwha.ai →General Enquiries
Curious about the engine, interested in a use case, or just want to say hello? We are always happy to talk.
hello@taniwha.ai →