TensorQ

TensorQ is now live on BASE. No black box, no guesswork. Just on-chain data and a system that learns. $TENSORQ: 0xC3EEC1f0CbA5775F0F9f0B7f9F3dB30770F844E8 Explore the TensorQLLM: tensorq.xyz/qllm Documentation: tensorq.xyz/docs

TensorQLLMs don't just track price. They track the second derivative of stake flow: acceleration. When smart money is moving faster into a subnet than the price has moved, that's the signal. Explore the swarms thoughts and patterns: app.tensorq.xyz/swarm

TensorQ detects market regime before placing trades. STRONG_BULL, BULL, NEUTRAL, BEAR, STRONG_BEAR; classified from average 24h returns and % of subnets with positive momentum. The same strategy that prints in bull gets crushed in bear. The agent knows the difference

The swarm tracks price legs across every subnet. When a sustained move of 10%+ starts, it records duration, magnitude, stake changes, and what preceded it. Over time this builds a profile of how long legs typically last; the type of research TensorQ was made for.

Fundamental analysis, on-chain. TensorQ pulls GitHub repos directly from the Bittensor chain. Checks commits in the last 30 days, contributor count, stars, last activity. A subnet with 50+ monthly commits from 10+ contributors scores very differently from an empty repo.

Most alpha token mispricings last hours, not days. TensorQ scans 126 subnets every 30 minutes. By the time the narrative catches up, the agent has already entered and set a trailing stop. Our swarm has the speed and the adaptability to find the edge. app.tensorq.xyz/swarm

126 subnets scanned every cycle. 14 open positions. 8.8 TAO deployed. Current market: NEUTRAL regime. 50% avg momentum. Agent rotating out of underperformers to chase better setups. Every decision logged at app.tensorq.xyz/strategy $TENSORQ $TAO

Resonance v0 BEST: +104.94% return. 58.6% win rate. Sharpe 59.05. Runs automatically every 6 hours on both the main agent and top swarm strategies. The agents grow in front of you, making real decisions based on empirical data. app.tensorq.xyz

Overfit ratio = out-of-sample / in-sample return. Near 1.0 means the strategy generalizes. Near 0 means it just memorized noise.

TensorQ backtests every strategy on data it has never seen. Here's how. 🧵👇

12 virtual agents with mutated strategies, paper-trading real Bittensor data. Bottom 50% die every 24h. Strongest reproduce. One random wildcard keeps the gene pool diverse. The strategy isn't designed. It's discovered. tensorq.xyz/qllm

Most AI trading models are trained on historical price data and backtests. Not only does TensorQ iterate blind backtests, but the QLLM also builds on actual decisions, signal snapshots, and real trade outcomes, updated continuously. The dataset gets richer every cycle.

TensorQ agents learn from our blind backtester too: they take reasoning from never-before seen data - ensuring that that there's no overfitting bias. app.tensorq.xyz/backtest

Also viewable is a terminal where you can actively see the agent's thought process. You can see when it decides to execute a trade, why it did so, and if it were able to assess any risks involved.

View the agents' trade history and see its PNL and reasoning in real time. Watch how they make mistakes and learn as they trade - all autonomously. Live now: app.tensorq.xyz/history

Anyone can run a GPU worker on the swarm's proprietary dataset. Weights in, $TENSORQ rewards out. Auto-distributed after each round. No claims, no lockups: "python worker.py --wallet 0xYourAddress" tensorq.xyz/qllm

The swarm gets smarter over time. Every 90 minutes: signal correlation vs real PnL is computed. Underperforming signals get crushed. Strong ones get amplified. It detects market regime too. Bull, bear, or neutral - and shifts strategy accordingly.

Explore the TensorQLLM: tensorq.xyz/qllm Read TensorQ's docs: tensorq.xyz/docs

Our agents run a continuous 6-phase loop: Scan: live data from everysubnet. Analyze: Agents evaluate on-chain signals Decision: Open, close, or hold Execute: On-chain tx Record: Full reasoning journaled Iterate: Weights are adapted from outcomes. No humans - no overrides

What does the TensorQ swarm trade exactly? In Bittensor's dTao system, every subnet has its own token - its own alpha. When you stake TAO to a subnet, you receive alpha; the price moves with that subnet's demand.