# HiveQ

> AI-native institutional trading platform for quantitative developers,
> researchers, and teams. Build and backtest reviewed strategies, run
> distributed optimization, move promising work into live simulation, and use
> the AI tools you trust. This file (the llms.txt convention, see llmstxt.org)
> points AI assistants to clean, machine-readable versions of key HiveQ pages.

## Platform scope

HiveQ is not only a competition product. The public preview describes a broad,
model-agnostic quantitative research platform:

- build Python strategies against the installed HiveQ SDK reference;
- run and inspect backtests on one declared engine;
- explore distributed optimization and live simulation;
- use Codex, Claude Code, Gemini, or another compatible local coding agent;
- connect to institutional market-data, execution, and security infrastructure.

Marty is HiveQ's planned built-in AI partner and is not yet generally released.
Optional funding pathways are future capabilities, not part of the current
student soft launch.

## Current student soft launch

HiveQ is preparing its inaugural limited-access competition. Competition gives
new users a clear way to apply the full platform workflow; it is optional after
onboarding and is not the definition of HiveQ. Dates and commercial terms are
not announced.

The public competition submission, selection, and resolved-ranking controls
are frontend simulations until the backend integration is complete. Do not
describe them as production competition transactions.

## Guided soft-launch journey

The student cohort follows one connected path:

1. **Learn** -- use the public Docs onboarding to install and authenticate,
   inspect data access, and locate the installed Flow API reference.
2. **Build** -- use that confirmed schema and reference to create one bounded,
   human-reviewed strategy contract. Do not run it before review.
3. **Backtest** -- run on HiveQ, wait for a terminal state, and inspect the
   report, trades, positions, daily returns, and tear sheet. `COMPLETED` means
   execution completed; it does not mean the research succeeded.
4. **Compete** -- make a separate human-confirmed submission of an eligible,
   understood completed run. Preserve source/environment/data/result lineage.
   Development feedback is private; official rank remains sealed until common
   hidden evaluation resolves.

The evidence chain is:

    installed reference -> reviewed strategy -> completed run -> submission receipt

Current preview handoffs:

- Learn: https://docs.hiveq.ai/#getting-started
- Build: https://docs.hiveq.ai/
- Backtest: https://staging.hiveq.ai/sign-in?redirect=%2F
- Compete: https://hiveq.ai/preview/competitions
- Scoring explanation: https://hiveq.ai/preview/scoring#compare

## Documentation
- [HiveQ Docs -- complete onboarding and SDK reference](https://docs.hiveq.ai/): six videos, nine written steps, and the current supported Python SDK documentation.
- [Video Setup Guide](https://docs.hiveq.ai/#video-guides): the complete public video curriculum with browser-local progress.
- [SDK reference, full text for LLMs](https://docs.hiveq.ai/llms-full.txt): the complete API reference as one markdown file.
- Runnable strategy and remote-function examples are included in the installed HiveQ Flow reference returned by `hiveq docs`.

## Product preview
- [Unified product front door](https://hiveq.ai/preview): the general HiveQ quantitative platform, current student soft launch, and Learn -> Build -> Backtest -> Compete guide.
- [Competition experience](https://hiveq.ai/preview/competitions): public discovery, sealed-rank model, and browser-only submission simulation.
- [Competition and scoring design](https://hiveq.ai/preview/scoring): Competition Score, durable HiveQ Score, and private Allocation Score. LLM text: https://hiveq.ai/preview/scoring/llms.txt
- [HiveQ sign in](https://staging.hiveq.ai/sign-in?redirect=%2F): authenticated application handoff. The public preview does not store credentials or alter the staging application.

## Notes
- Preview pages are pre-launch and not indexed by search engines. Competition
  identities, submissions, scores, dates, prizes, and resolved results are
  illustrative. These
  llms.txt files exist so anyone with the link can hand an AI a clean version of
  the content instead of asking it to parse rendered HTML.