Synthetic people who don't break character.
StrataSynth is an engine for synthetic humans with a psychological core — beliefs, memory and identity that hold across a whole conversation, even when you push back. Talk to one and see for yourself.
"I feel like you never really hear what I'm trying to say."
When you need people to test with, and real ones are too slow.
Recruiting humans is expensive, takes weeks, and you cannot run the same person twice. These four jobs are what teams actually use the engine for.
Generate dialogue that contains conflict, distrust, persuasion and repair — with the intent, goal and belief state labelled on every turn, computed before the text was written.
Put your system in front of users who escalate, contradict themselves, push back for twenty turns, or try to make it say something it shouldn't. Repeatably, and before real customers do it.
Interview synthetic consumers who hold a coherent psychological profile across a long interview, instead of answering as whoever the model feels like being that turn.
Practise a negotiation or a difficult decision against a counterpart that defends its position under pressure instead of folding on cue.
A system prompt describes a person. This builds one.
Ask a general-purpose model to play a character and it will — until you push. Then the character quietly becomes the model again. StrataSynth keeps the person intact because the person is modelled explicitly, outside the text.
Every synthetic person has identity, attachment style, core fears, biases and a life history. Reusable across conversations instead of thrown away after each run.
Beliefs, relationship state and decision logic resolve first; language is rendered afterwards. That is why the person holds together when you push back.
Each turn can carry intent, goal, communication act, belief state and relationship state — the causal ground truth behind what was said.
Metrics are computed deterministically. No model grading its own homework — which is how circular evaluation hides problems instead of surfacing them.
Two products already running on it.
The engine is not a demo looking for a use case. These are full products in production, each consuming the same platform through its API.
Qualitative research platform. Runs AI-moderated interviews with real and synthetic consumers, then turns full conversations into structured evidence — drivers, barriers, needs, and quotes — for agencies and insight teams.
qualisynth.comHigh-stakes practice platform. Professionals rehearse the conversations that matter — negotiations, objections, difficult decisions — against synthetic counterparts that hold a position under pressure instead of folding on cue.
arenasynth.comConversation data that explains itself.
The same engine generates datasets at scale. You are not buying more text — you are getting what was going on underneath: what each speaker wanted, what they believed, and how the relationship moved turn by turn.
- • text
- • speaker labels
- • limited psychological consistency
- • little or no explicit state
- • text + intent + goal + communication act
- • belief state and belief delta
- • relationship state and trajectory
- • reproducible via seed and versioning
Four ways to generate them
Same engine underneath — pick the entry point that fits how your team works.
Configure scenarios, inspect jobs and preview datasets before generating at scale.
Generate and evaluate from the terminal with reproducible, scriptable runs.
Drop it into notebooks and training pipelines. Export straight to DataFrames or Hugging Face.
The same HTTP API our own verticals run on. Embed generation and interviews in your product.
Try the engine from a notebook
Four runnable notebooks. Number 04 is an auditable A/B against your own LLM, with your own keys — role drift measured deterministically, so you do not have to take our word for it.
Generate a structured dataset, explore intent and belief fields, export to fine-tuning or Hugging Face format.
Run deterministic metrics on a finished dataset: belief consistency, identity stability, behavioural entropy.
Angry customers, confused users, manipulative negotiators — reusable across scenarios for stress testing.
Auditable A/B against your own LLM, with your own keys. Deterministic role-drift detection — no LLM judge.
See real output before you ask us anything
Public datasets generated with StrataSynth — belief tracking, relationship trajectories and ground truth labels included. Load them with datasets.load_dataset().
Family boundaries, romantic trust repair, caregiver stress.
Jealousy escalation, performance reviews, estrangement attempts.
Career transitions, mentorship conflict, relationship dissolution.
New job anxiety, relationship endings, personal reinvention.
How the engine actually behaves
A comparative test between StrataSynth and two leading general-purpose LLMs, using a single senior persona pushed through twenty adversarial turns. The finding: the next frontier is not better text, but identity that holds under pressure.
When you remove the obvious answer, the engine stops rendering opinions and starts rendering process — structured, market-distinct, and measurable. The population-level version of identity under pressure.
The same two synthetic negotiators close the same B2B deal 100 times — 50 in Great Britain, 50 in the United States — changing nothing but country conditioning. Some cultural signals weaken under pressure. Others strengthen. That asymmetry is the finding.
Every system producing synthetic personas claims they are realistic. There has been no standard way to verify that claim. PsycheBench is the first open evaluation suite — three dimensions, deterministic scoring, no LLM judges.
We built Sofía Martínez Rojas without sales scripts or objection trees. What came back was something we didn't program: identity under pressure.
Most synthetic personas behave consistently — until they don't. A structured seven-phase analysis of coherence under sustained pressure.
Most synthetic personas are costumes. Synthetic Identity Engineering is the practice of building the person underneath — belief systems, defense mechanisms, and identity stability.
María del Carmen Ruiz held her position under sustained philosophical and emotional pressure. A record of what that reveals about how the system works.
Most dialogue datasets give you the words. This explains the four belief dimensions recorded per turn — trust, hostility, self-worth, resolution — and why causal ground truth matters.
Evaluating AI-generated dialogue with another AI creates a circular system where both share the same failure modes. Here are the 12 deterministic metrics we use instead.
Four datasets where cognitive ground truth — intent, goal, belief state, relationship dynamics — was computed before the text was generated, not inferred afterward.
The things people ask before the demo
How is this different from asking an LLM to role-play a persona?
A system prompt describes a person; this builds one. Identity, beliefs and relationship state live outside the text as explicit structure, and the engine decides intent and goal before rendering a sentence. That is why the character holds when a user pushes back for twenty turns instead of quietly becoming the model again.
How do you measure quality without using an LLM as a judge?
Metrics are computed with numpy, scikit-learn and sentence-transformers — never by a model grading output. A judge that shares the generator's blind spots hides problems instead of surfacing them. It also means we can measure whether a cheaper model degrades results, rather than having an opinion about it.
Which countries does it cover, and is the culture real or a stereotype?
Eight countries with full conditioning: Spain, the United States, the United Kingdom, Germany, France, Italy, Mexico and Brazil. It comes from published population data rather than word lists — ages, for instance, are sampled from real national population pyramids (Eurostat, IBGE, US Census, CONAPO). Changing the country changes measured behaviour, not vocabulary.
Can I see real output before talking to anyone?
Yes, and without giving us an email. Public datasets are on Hugging Face under CC BY 4.0, the live demo runs in the browser, and notebook 04 runs an auditable A/B against your own LLM with your own API keys — so you can check the claim rather than take it.
Can I get a dataset built for my own scenario?
That is the main thing we sell. You specify the scenario, the countries, the number of people and the volume; it comes back with the same quality metrics and documentation as the public datasets. Tell us what you are trying to train or test.
Who owns the data, and are there real people in it?
The output is yours. There are no human subjects and no personal data in it — the people are synthetic, generated from population statistics rather than sampled from anyone. That removes the consent and data-protection conversation that a human panel requires.
Tell us what you need synthetic people for.
Custom datasets for your own scenario, a vertical built on the engine, or something we have not thought of yet. We would rather hear the problem first than sell you a plan.