Competitive intelligence
Reads the signals inside released games and turns them into clear views of product direction, design choices, monetization, and technical strategy.
Independent product · Live
Game competitive intelligence
See beyond the charts. Understand how a game was built.
HeyCici helps game teams look inside shipped products—not to copy them, but to understand the decisions, systems, and patterns that made them work.
✓Running as a real production platform
✓Turns product evidence into clear competitive briefs
✓Provides the foundation behind AI-assisted game reconstruction
~100games analyzed
~2.5Mcore-game functions indexed
ProductionBI and data platform status
Hoursfrom a shipped build to structured evidence
Why HeyCici exists
The market tells you what is popular. We wanted to understand why.
Downloads and revenue only describe the surface. The useful questions live deeper: what changed between releases, how a game earns money, how its levels evolve, which systems are reused, and where the team placed its engineering effort. HeyCici was built to turn those questions into evidence a game team can actually use.
One product, three connected systems
AI-assisted reconstruction
Uses grounded project evidence to help rebuild understandable game systems and shorten the distance from research to a working prototype.
Data platform
Keeps evidence, relationships, reports, and workflows in one operational product so each new analysis makes the next one stronger.
How the system works
A repeatable path from a shipped build to competitive knowledge.
Ingest the released product
A game build enters the pipeline together with its version identity. The system supports the evidence available from common Unity and Cocos release formats without requiring the original source project.
Extract and normalize evidence
Code structures, calls, configurations, assets, levels and technical fingerprints are processed into a consistent model instead of remaining as disconnected files.
Analyze with rules and grounded AI
Deterministic services handle facts such as SDK signatures and version relationships. AI is used for synthesis—reading the evidence, connecting patterns and producing briefs with traceable support.
Deliver insight through product surfaces
Results become searchable intelligence, reports, APIs, dashboards and inputs for the AI Re-coder, allowing product, design and engineering teams to work from the same evidence layer.
What we have already made possible
From a released game to knowledge a team can act on.
Follow how a product evolves
HeyCici can compare releases and reveal where a team changed direction—whether that shift happened in gameplay, difficulty, monetization, or live operations.
Read the business model from the product
It can surface the advertising, analytics, payment, offers, and progression choices embedded in the game, then translate them into language product teams understand.
Recognize the architecture behind a portfolio
Shared frameworks and recurring patterns become a kind of technical DNA, showing how studios reuse foundations and where they invest in control or speed.
Turn assets and levels into design evidence
Instead of leaving extracted data as raw files, HeyCici connects it back to the player experience so designers can study structure, content, and intent.
A real report example
The eyes and the AI leverage in the game industry
This public research note shows the kind of deliverable HeyCici can produce: a readable argument supported by evidence recovered from shipped games—from level structure and monetization to SDK usage and architectural patterns.
BI ToolsGame IntelligenceAICompetitive Analysis
Read the full HeyCici report ↗Level 760—visualized directly from structured game data.Selected verified evidence
The system does not stop at metadata. It reaches the decisions inside the product.
The following figures come from selected production analyses. Game and studio names are intentionally withheld for a public portfolio.
levels mapped as a design curve
A complete progression revealed how difficulty was distributed and how board complexity changed over time. A version comparison also exposed a major rewrite in core board-generation behavior.
real IAP configurations recovered
Bundles, currencies, boosters, remove-ads offers and seasonal products could be read as an actual monetization catalog—not inferred from screenshots or market estimates.
shared-platform components reconstructed
A released product exposed the structure of a reusable casual-game framework spanning live operations, levels, boosters, payments, player state and offers. The result is readable and compilable; behavioral equivalence remains a separate validation problem.
assets organized from one product
Images, prefabs, animations, shaders, audio and configuration objects were indexed and connected. Image fingerprints make art changes and reuse measurable across releases.
levels restored from raw data
Grid coordinates, directions, rewards and content structure were recovered deeply enough to recreate the visual intent of levels and analyze a full content library instead of isolated screenshots.
relationships across game portfolios
Shared namespaces, framework patterns and asset fingerprints reveal which products inherit the same foundations, where code was forked, and how studios trade speed for control.
Who this changes work for
Game Design
Study real difficulty curves, progression, economies and level structures before committing months to a hypothesis.
Engineering
Learn from frameworks and architectural patterns that already survived production, while seeing their trade-offs rather than copying blindly.
Product & UA
Connect monetization, SDK usage, offers and release changes to a clearer picture of how a product is operated.
Art & Content
Compare asset libraries and visual changes across versions with evidence instead of manual cataloguing.
Product architecture
Built as a production system, not a one-off reverse-engineering script.
HeyCici separates extraction, evidence storage, intelligence and delivery so each layer can evolve independently. New games enrich the corpus; shared frameworks and recurring patterns make later analyses faster and more useful.
Automated workers
Version-aware pipelines coordinate unpacking, static analysis, asset processing and large batch workloads.
Unified data layer
Structured records, object storage, relationships and searchable context preserve the chain from a conclusion back to its source.
Rules + AI services
Deterministic detectors, comparison services and grounded AI analysis turn raw evidence into product knowledge.
SaaS, APIs and workflows
Dashboards, CMS, reports, integrations and AI tool interfaces expose the same intelligence to different teams.
PythonFastAPIPostgreSQLMinIORedisPrefectDockerGCPUnityStatic & runtime analysis
How we use AI
AI should amplify judgment, not manufacture certainty.
HeyCici is designed around evidence. When a result can be determined by rules, we use rules. When AI interprets the evidence, the conclusion stays connected to its source. If the system cannot support a claim, it should say so. The goal is not an impressive answer—it is a trustworthy one.
I designed and built HeyCici end to end—from product direction and system architecture to analysis pipelines, backend services, AI workflows, deployment, and production operation.heycici.
HeyCici is already more than an idea.
The foundation is running, the product is learning from real game evidence, and the next step is turning that depth into a tool more teams can use.