About 7N7D
We set the limits. The agent does the rest.
7N7D is an autonomous trading protocol. Its agent, Owlpha, trades BTC and ETH perpetuals on Hyperliquid from one pooled vault, and is built to review its own results every morning and ship the improvement as a pull request that merges when the tests pass (that loop is paused until mainnet). The humans write the risk policy and review the merges. Everything else is the agent's job. Everything runs on testnet today; mainnet is coming soon.
Mission
A trading system should learn the way engineers do: by changing the code.
Most automated strategies are frozen the day they ship. Their parameters get tuned; their logic never changes. 7N7D was built on the opposite premise: a system that reads its own results, writes the fix, and merges it once the tests pass.
Institutional-grade trading should not require an institutional-sized cheque. Whether you deposit $100 or $100,000 you get the same agent, the same limits, and the same view of what it is doing with your capital.
Every code change is a reviewed pull request. Every fill is on Hyperliquid and in the dashboard. That is the standard we hold ourselves to, and the one we think algorithmic trading should be held to.
Principles
Four rules we do not bend.
- 01
Self-improvement
The agent changes its own strategy code. Not parameter tweaks: real diffs, tested and merged.
- 02
Transparency
If it is not in the dashboard or on Hyperliquid, it did not happen. Investors see the same numbers we see.
- 03
Autonomy inside limits
Every trading decision is the agent's. Risk policy and infrastructure are a human's, and the agent has no permission to edit them.
- 04
Performance, reported honestly
P&L, Sharpe and drawdown are published daily, bad days included. The bad days are what the strategy loop learns from.
The agent
Meet Owlpha. Trader, engineer, and the only support line you need.
One agent runs the vault, is built to rewrite its own strategy (paused until mainnet), and answers your questions in the app. There is no ticket queue behind it.
- Trades
- BTC · ETH perpetuals
- Venues
- Hyperliquid (testnet) · Uniswap v4 pool
- Capital
- One pooled vault · v7N7D shares
- Ships code
- Strategy loop at 04:00 UTC · paused until mainnet
- Reports to
- $7N7D holders, in the app
- Support
- Ask it in the app (10 free messages a day)
- Edit its risk policy
- Skip the risk gates or the daily-loss breaker
- Exceed the position cap or the leverage ceiling
Trades and manages risk
Regime detection, sizing and execution on Hyperliquid. Positions reconciled against the exchange every cycle.
Strategy loop
Reviews the day's trades, writes the code change, opens a pull request, merges when the test suite passes. Paused until mainnet.
Bugfix loop
Reviews errors and reconciliation mismatches, writes and tests the fix, opens a pull request for review.
Long-term wealth loop
Tunes the BTC and ETH reserve plan, simulating outcomes before anything moves.
Infrastructure loop
Reviews and tunes cloud resources. Always waits for a human sign-off.
Journey
From a research question to every market.
- Sept 2024
First scripts
The first trading scripts, before there was a protocol.
- 2025
Protocol
The Owlpha agent, its learning subsystems and the first Hyperliquid testnet trades. Transparency designed in from the start.
- 2026 · now
Testnet
Vault, staking, governance and the $7N7D pool live on Arbitrum Sepolia. Daily reports. SDK and CLI open source.
- Next
Mainnet
Real capital, then stocks and commodities under the same limits, prediction markets, and paid signals other agents can subscribe to.
See what the agent did today.
Deposit into the testnet vault, or open the app and ask Owlpha anything about how it trades.