01 / AI
Persistent AI conversations
DINAV maintains research conversations through its Core API and configured OpenAI-compatible model endpoint.
ASR Products · DINAV
Quantitative research infrastructure, built around evidence.
DINAV is Alpha Stochastic Research’s flagship quantitative-finance terminal for governed AI research, market intelligence, point-in-time analysis and deterministic finance computation. Its production backend combines persistent AI conversations, file-assisted research, market/news/event access, point-in-time replay, scholarly metadata search, portfolio context, signed artifacts and bounded deterministic finance calculations.
ALPHA STOCHASTIC RESEARCH / FLAGSHIP PRODUCTVerified capabilities
What DINAV actually provides.
These capabilities are taken from the production DINAV repository and architecture. The page does not substitute mock features when a service or provider is unavailable.
02 / FILES
File-assisted research
Uploaded research material can be used as context inside the governed AI research workflow.
03 / MARKETS
Market, news & event access
The Core API queries DINAV market intelligence stored in Supabase and news ingested from approved providers.
04 / PIT
Point-in-time replay
Availability time is preserved separately from publication or observation time so historical replay can filter on what was actually available.
05 / PAPERS
Scholarly metadata search
DINAV connects scholarly metadata workflows to arXiv and Crossref rather than fabricating papers or identifiers.
06 / PORTFOLIO
Portfolio context
Portfolio context is part of the production Core API capability set for research and analysis.
07 / ARTIFACTS
Signed research artifacts
DINAV can register signed artifacts backed by private storage and its artifact registry.
08 / COMPUTE
Deterministic finance compute
DINAV Compute exposes bounded finance calculations. Arbitrary public Python execution remains disabled unless a separately isolated runtime is configured.
09 / ACCESS
Authentication & entitlements
Identity, usage and product access are governed through Supabase Auth, row-level security and DINAV entitlement logic.
Production architecture
One product with explicit runtime boundaries.
DINAV separates AI orchestration, ingestion, deterministic computation and the data authority. This keeps provider secrets and privileged infrastructure outside the browser.
Read architecture →DINAV Core API
Authentication, entitlements, AI orchestration, conversations, files, market intelligence, scholarly metadata, portfolio context and artifacts.
DINAV News Worker
Governed ingestion of approved providers into the point-in-time news store.
DINAV Compute
Constrained deterministic finance operations with no arbitrary public Python shell.
Supabase data authority
Source of truth for identity, entitlements, usage, workspaces and DINAV market intelligence.
Safety invariants
Missing data stays missing.
DINAV’s repository defines explicit boundaries for market data, papers, model state, computed results and secrets.
No fabricated provider state.
DINAV does not fabricate prices, headlines, provider status, papers, identifiers, model state or computed results. Missing feeds remain explicit no-data or configuration states.
Secrets remain server-side.
Provider secrets, model-provider keys, Supabase service-role keys and compute-worker tokens are not sent to the browser.
DINAV is the flagship ASR quantitative-finance terminal.
Open the DINAV platform directly from ASR Products.