Skip to main content

Intelligence infrastructure for the game industry

FAQ

What Iridae is, how it works, where it fits, and how teams stay in control. Need a specific answer for your team? Get in touch.

Loading questions...

Core model

What is Iridae?

Iridae is intelligence infrastructure for the game industry. It connects your studio's tools and data sources into one system, catches decisions misaligned with outcomes, and corrects them in real time. It shows up inside the tools your team already uses, and nothing runs without approval.

Read the product overview

Core model

What are the four core systems?

Latent Spark is a living probabilistic atlas of the game market. Subtle Beacon handles Bayesian experimentation and preference learning. Patient Cartographer models strategic planning as movement through a changing resource landscape. Anchored Horizon provides dynamic commercial forecasting. All four share an uncertainty-aware reasoning layer so they compound each other's outputs.

Explore the systems

Core model

What is Latent Spark?

Latent Spark gives every game a structured, continuously updated probabilistic profile combining mechanics, identity, player response, commercial signals, and market context. It represents what's known, what's likely, and how reliable that confidence is. It powers concept evaluation, comp selection, and market mapping across the rest of the stack.

Read about Latent Spark

Core model

What is Subtle Beacon?

Subtle Beacon is a Bayesian experimentation and preference-learning system. It treats player preferences and experiment outcomes as evolving probability distributions, so teams can evaluate changes while data is still arriving, adapt traffic allocation, and measure decision-readiness rather than waiting for a test to feel "finished."

Read about Subtle Beacon

Core model

What is Patient Cartographer?

Patient Cartographer models strategic planning as movement through a changing resource landscape. It explores many plausible tactical sequences, identifies which reach your objectives given real constraints (budget, team capacity, production progress, market position), and distinguishes paths that look good on paper from those that hold up under execution risk and timing.

Read about Patient Cartographer

Core model

What is Anchored Horizon?

Anchored Horizon is a dynamic commercial forecasting system. It retrieves comparable games algorithmically, builds structured priors, and produces probabilistic forecasts (not point estimates) that update as your concept evolves. It handles scenario-native queries: how do outcomes change with different product directions, pricing, launch timing, or scope?

Read about Anchored Horizon

Core model

What research does Iridae publish?

Iridae publishes research across three pillars: intelligence-in-flow (how humans and models share reasoning), behavioral intelligence (mapping player cognition to market outcomes), and computational intelligence architecture (real-time probabilistic reasoning). Published research includes Grafial (a language for reasoning about uncertain relationships), RVGN (reasoning with incomplete adjacency), SATE (online topic segmentation), and the Spectral-Chaos Transformer (scenario forecasting with structure and uncertainty).

Explore R&D

Core model

How technical is Iridae under the hood?

Very. Iridae is built on frontier research in uncertainty, behavioral intelligence, and live decision-making. The engineering layer enforces versioned artifacts, rigorous testing, tenant isolation, and reproducible workflows. The four core systems share an uncertainty-aware reasoning language so they compound rather than conflict.

Read the technical overview

Core model

Why doesn't this already exist?

Three reasons. First, the problem is illegible from the outside: it's a sequential decision process where each call constrains the next, and the waste compounds invisibly. Second, the incumbents are subsidized by the inefficiency: platforms, ad networks, and middleware providers earn more when studios spend more to find fit. Third, the solution requires frontier research, not better tooling. Modeling the relationship between a game and its market under radical uncertainty, with non-stationary data, is a science problem with commercial implications. Nobody built it because nobody had the cross-domain team or the research to do it.

Read the product overview

Core model

What is GMRL?

GMRL (Game Market Research Lab) is a separate Iridae product that turns your pitch, design docs, or Steam page into a clear market read with drop-in visuals. It's useful for publisher conversations, greenlight decks, and positioning validation. You can use GMRL independently of the Iridae platform.

Visit GMRL

Category Comparisons

Is Iridae a publisher?

No. Iridae does not take ownership of your studio or IP. It gives your team stronger operating discipline so you can go to market on better terms, with or without a publisher.

Read: Iridae vs Publishing

Fit and scope

Who is Iridae built for?

Teams that own decisions and carry consequences: studio founders, game directors, producers, engineering leads, creative directors, and marketing/growth leads. Iridae is built for the cross-functional reality of game development where decisions touch product, production, creative, engineering, and go-to-market simultaneously.

Explore role-based fit

Fit and scope

Can a solo developer use Iridae?

Yes. Solo devs make all the same decisions as a full team, just alone and disconnected from market signal. Iridae watches the market side while you build: continuous reads on your store page, community, playtest signals, and player behavior. It surfaces the next move scoped to what one person can ship, and holds the thread across decisions so context doesn't evaporate between sprints.

Read: Iridae for solo developers

Fit and scope

Does Iridae replace our existing tools?

No. Iridae works through your existing stack. It integrates with project management tools (Linear, Jira, Trello, ClickUp, Asana, HacknPlan, Codecks), team chat (Slack, Discord, Teams), docs (Notion, Confluence, Google Workspace), analytics (GameAnalytics, Amplitude, Mixpanel), version control (Git, Perforce), and more. It also supports REST API, GraphQL, and n8n/Zapier for custom workflows.

See all integrations

Fit and scope

Is Iridae only useful in live ops?

No. Iridae is built for decisions across the full lifecycle: first concept (validating direction before committing resources), vertical slice and production (keeping creative vision aligned with market reality), pre-launch and launch (building signal, not just hype), and live ops (making the update loop compound instead of reset).

Explore fit by development phase

Fit and scope

Is Iridae only for large studios?

No. Iridae supports solo developers, small indie studios, mid-size studios, large studios and publishers, and accelerators/incubators/funds. Fit is driven by whether you have real decisions under real constraints, not by headcount.

Explore fit by studio size

Fit and scope

Can Iridae be used outside games?

Yes. The core technology is not game-specific. Product priorities are shaped by game-industry needs, but Iridae stays engaged with a limited number of non-game pilots where the decision patterns overlap: uncertain markets, cross-functional execution, and outcomes that compound over time.

Discuss a non-game pilot

Fit and scope

Who is behind Iridae?

Iridae is based in Copenhagen, Denmark (Iridae ApS). The founding team includes Charles Hinshaw (CEO, early Unity employee #15), Emil Johansen (CTO, early Unity engineer #22), Ghazaleh Mobkie (COO, early Trustpilot employee #38, former Dixa), Soren Lass (Developer Relations, former Ubisoft brand and marketing director), and Alessandro Canossa (ML & Research, applied researcher on player behavioral models).

Meet the team

Fit and scope

Where is Iridae based?

Iridae ApS is based in Kongens Lyngby, Denmark, in the Copenhagen area. The company is a Danish ApS (limited company) and operates under EU/EEA data protection regulation.

About Iridae

Governance

Do you provide a Data Processing Addendum (DPA)?

Yes. Iridae publishes a DPA covering controller/processor responsibilities, security measures, subprocessor management, cross-border transfer safeguards (Standard Contractual Clauses), and deletion obligations (customer content deleted or returned within 60 days after termination).

Read the DPA

Governance

Is customer content used for model training?

Per the AI Policy, customer content may be used to improve the service unless a written no-training agreement is in place. Iridae uses privacy-preserving, federated-learning-style approaches where learning is captured as model updates or weights rather than exposing raw customer content across accounts. This is not a prompt-in/public-recall-out workflow. If you need a no-training guarantee, request a written agreement.

Read the AI Policy

Governance

Where is our data stored?

Personal data is primarily located in the EU/EEA. Where transfers outside the EU/EEA are necessary (e.g. to subprocessors), Iridae relies on Standard Contractual Clauses. Iridae does not sell personal data. Controller-side data is retained as needed for stated purposes; processor-side customer content is deleted or returned within 60 days after contract termination.

Read the Privacy Policy

Governance

What subprocessors does Iridae use?

Iridae uses a combination of first-party models and third-party providers. Infrastructure: Amazon Web Services, Microsoft Azure, and NVIDIA for hosting and compute. Third-party model providers: OpenAI and Anthropic (used selectively). All subprocessors have Data Processing Agreements with Standard Contractual Clauses. The full list with roles, data categories, and locations is published.

View the subprocessor list

Governance

Does Iridae use its own AI models?

Yes. Iridae uses a combination of first-party models (developed internally) and third-party models. Requests may be routed between systems for quality, safety, latency, or availability. Some requests are processed or reviewed by humans for input preparation, output quality control, or safety, subject to confidentiality obligations.

Read the AI Policy

Commercial

Can we license individual subsystems?

No. The four systems (Latent Spark, Subtle Beacon, Patient Cartographer, Anchored Horizon) are designed to compound each other's outputs. They share an uncertainty-aware reasoning layer and produce better results together than in isolation. Iridae is offered as an integrated platform.

Ask about pilot engagement

Commercial

Is Iridae generally available?

Not yet. Iridae is currently offered through invite-only pilots. Get in touch and tell us about one decision your team is facing. That's the fastest way to evaluate fit and get into the pilot queue.

Get in touch

Commercial

How much does Iridae cost?

Iridae does not publish flat self-serve pricing during the invite-only pilot phase. Commercial structure is scoped with each team based on pilot scope and operational needs. Reach out to discuss what a pilot looks like for your studio.

Discuss pricing and scope

Commercial

How do pilots work?

Start with a short call and one active decision. We map scope, constraints, and what an initial pilot should prove. Pilots are scoped around real decisions with real stakes, not sandbox demos. We support a limited number of non-game pilots where the decision patterns overlap.

Start a pilot conversation

Commercial

How do we get started?

Get in touch and bring one concrete decision where the tradeoffs and downstream commitments are unclear. That's the fastest way to evaluate fit. We'll show you where Iridae would plug in and what a pilot would look like for your team.

Get in touch