dew
Foundation model for under-16 users

Why expect kids to use AI responsibly, why not build AI which behaves responsibly?

Most AI is built for adults. When children use it, we add filters, limits, and parental controls — but children are not “smaller adults.” They are still developing the reasoning, creativity, and perception of world that AI can so easily weaken. dew is a child-facing AI companion, that kids (under 16) can turn to — to learn, create, share, discuss, and find guidance — while protecting the very thinking that makes learning meaningful. Grounded in the Socratic method, it guides rather than hands over answers — affordable, open-source and built as a public good.

See it in action

A foundation model that is child aware.

A look at dew engaging with a child, guiding and coaching them through academic, social, and emotional queries.

Now playing
Socratic walkthrough
A supportive chat
Runs offline, on any device
For child

More than a coach. A companion.

Learn by thinking, not copying

Ask anything, any time. dew never rushes to answer. It helps a child build the habit of thinking, not just the habit of prompting.

Share & discuss, freely

A safe space to talk through ideas, curiosities, and worries — dew listens, responds at their level, and keeps the conversation non-judgemental.

Speaks a child's language

Pre-trained on billions of educational tokens — K-12 textbooks, curriculum guides, and public-domain literature, so its vocabulary and pacing fit the student.

Screenshot / photo — student using DEW
Screenshot / photo — parent & child
For parents

Safeguarding without surveillance.

Alerts for serious risk

If a child shows signs of suicidal thoughts or violent ideation, parents are alerted immediately.

Private by design

No ads, no data harvesting. Inputs are stripped of personal information before the model ever sees them.

A bi-weekly wellbeing report

Every two weeks, parents receive a summary of their child's mental health, areas of interest, areas of concern, and learning progress.

For educators

AI you can finally allow in class.

Protects academic integrity

It won't generate passive essays or final answers, so assignments measure real understanding again.

Runs on what you already have

A light 3B model runs on standard lab desktops, tablets, or fully offline — no hardware upgrades.

On-prem & open-source

No student data leaves your institution. Every component can be inspected, customized, and deployed on infrastructure you control.

Screenshot — teacher dashboard
How DEW teaches

The Socratic method.

dew isn't adult AI with parental controls added later — it's developmentally aware by design. At its core is the Socratic method: a cooperative, question-and-answer style of teaching. Instead of providing answers, dew guides children to discover truths on their own — asking probing, open-ended questions that challenge assumptions, reveal contradictions, and build deep, analytical thinking. It knows when to answer, when to ask, when to simplify, when to encourage, and when to direct a child toward a trusted adult.

The kinds of questions it asks

Clarifying

“What exactly do you mean by that?” · “Can you give me an example?”

Probing assumptions

“How did you arrive at that?” · “What if this were the case?”

Exploring evidence

“Why do you believe this evidence is reliable?”

Examining perspectives

“How might someone on the opposing side view this?”

Testing consequences

“If your argument holds true, what logically follows?”

The result

The student becomes the active participant — thinking for themselves and building understanding that lasts.

How we built it ?

Pre-trained on billions of educational tokens — K-12 textbooks, curriculum guides, and public-domain literature — so its vocabulary and pacing fit the student, not a corporate user.

By the numbers
3B
Parameters — a light footprint tuned end-to-end, not a prompt wrapper on a corporate API.
<16
Age below which DEW hard-refuses to dump finished answers, guiding Socratically instead.
−∞
Logit clamp on forbidden concept classes — the network is physically incapable of expressing them.
Apache 2.0
Weights, code & datasets — white-label, self-host, and modify with zero restrictions.

Training pipeline

01 · Continued Pre-Training (CPT)

Initializes from a clean 3B baseline, then deep-trains on billions of educational tokens — K-12 textbooks, peer-reviewed curriculum guides, and public-domain literature — shifting the core language structure to age-appropriate vocabulary.

02 · Direct Preference Optimization (DPO)

Locks in the Socratic process over thousands of "chosen vs. rejected" pairs — chosen: multi-turn guiding inquiries, contextual hints, conceptual milestones; rejected: direct answers, final sums, passive essay generation.

03 · Logits-Level Security

A custom runtime LogitsProcessor identifies forbidden conceptual classes and clamps their generation probability to −∞ — bypassing brittle system-text blocks entirely.

Research

DEW: Reclaiming Classroom Thinking in the AI Era

By the DEW Team · June 2026

Our mission brief and technical architecture — how the Socratic method is built into the model's weights, the logits-level guardrails that keep it safe, and why we ground all of it in Responsible AI.

Read the paper ↗
See all research papers →
Our vision

It will be free forever

A child's mind should never be a business model. We intend to keep DEW free, forever — never gated behind a paywall, never sold. Our code, datasets, and methods stay open, always, under Apache 2.0, so anyone can inspect, host, and improve it.

This is how we deliver on our purpose of Responsible AI: not as a compliance exercise, but as the engineering philosophy that guides every model we build.

FAQ
Is DEW really free?+
Yes — free, forever. dew is a public good. We will never gate it behind a paywall, charge schools, or sell access to a child's learning.
What ages is DEW built for?+
dew is tuned for K-12 learners. For students under 16 it refuses to hand over finished answers, guiding them with Socratic questions instead; vocabulary and pacing adapt to grade level automatically.
Can we run DEW on our own infrastructure?+
Yes. dew is a light 3B model you can host entirely in-house via vLLM or Ollama — on standard lab desktops, tablets, or even fully offline. Student data never has to leave your network.
Is it open-source? What's the licence?+
Fully open under Apache 2.0 — code, datasets, and methods. You can inspect exactly how it works, white-label it, and customise the codebase.
How does DEW keep student data private?+
Inputs are sanitized of names, numbers, and other PII by design. There are no ads and no data harvesting, and zero student metadata leaks to third parties.
Get involved

We are open to partner

dew is free and open-source. We grow it with partners who bring it to schools, fund it through CSR, or collaborate on research.

Get in touch with the dew team at dew@rainwaterlabs.com