Financial reasoning shouldn't be a black box.
CHIAKOR is an explainable, knowledge-graph-based reasoning platform. It exists to make market cause-and-effect inspectable — so a recommendation is something you can follow, question and trust, powered by the frozen FinKG engine.
Opaque models, unaccountable calls
Most financial models output a number and hide the reasoning. When you can't see why, you can't audit, challenge or learn from it — and you certainly can't defend the decision.
The explanation should be the reasoning
CHIAKOR was built on a simple conviction: a recommendation is only trustworthy if its reasoning is explicit. So the explanation isn't generated after the fact — it is the computation itself.
A world where every financial decision carries its evidence with it — transparent, traceable, and open to inspection by the people it affects.
Principles, not slogans
Explainability
Every output decomposes into signed, weighted paths. Nothing is asserted without its evidence.
Transparency
Provenance travels with every edge and every path — you always know where a number came from.
Trust
Point-in-time correctness, honest limits and a demonstrator posture: we never overclaim.
A frozen engine behind a clean platform
CHIAKOR separates a deterministic reasoning engine from the platform that serves it. The engine is frozen and additive-only; the platform adds identity, APIs, observability and this interface.
- The engine never depends on the platform — reasoning stays pure.
- Language models are pluggable behind a vendor-neutral seam.
- Every layer is independently testable and observable.
Next.js/v1frozenpoint-in-timeBuilt in deliberate phases
The engine and platform were built additively — each phase frozen before the next began.
- Phase 1–2
Knowledge graph & reasoning core
The canonical graph, sign-algebra reasoning, RDF layer and evidence paths.
- Phase 3
Temporal, causal & portfolio intelligence
Point-in-time correctness, backtesting, calibration, portfolio and risk.
- Phase 4
Platform, identity & serving
REST API, JWT/RBAC, Postgres, observability, LLM seam and deployment.
- Phase 5
Product interface
The workspace, dashboards and this CHIAKOR experience.
- 6Phase 6
Continuous learning in production
PlannedLive outcome feedback and adaptive calibration — not yet released.
The architectural authorship of CHIAKOR.