Choose your moment
Pin an Ethereum block and verify its hash. Record funding, approvals and every local override.
THE ONCHAIN EXPERIMENT LAB
One moment.
More possibilities.
A time machine for onchain decisions. Return to a known state, try another action, and see what actually changes.
Open source. Local execution. Every outcome leaves a trace.
SOURCE / ETHEREUMONE STATE → FOUR PATHS
01 / EXPERIMENT CONSOLE
Explore the paths
before you choose one.
Compare Uniswap actions from the same archived Ethereum state. A successful transaction can still exceed your limits. Read the outcome and the reason together.
Four paths. One starting point.
Snapshot observations matchREAD THE TRADEOFF
Open the flight recorder
EXACT RAW UNITS, NO PNL CLAIMSPinned state, overrides & report integrity
Funding and approval receipts — separate setup gas
What this does not tell you
These transactions ran on a local fork of archived Ethereum state. They do not predict future prices, gas, MEV or market reactions. Artificial funding and impersonation are disclosed. A hash checks integrity, not truth. Token changes are not portfolio profit.
Set your next experiment
RUN LOCALLY · IMPORT HEREChoose constraints to build a command for your local runner. Run it on your computer, then bring the resulting JSON back to the console. This website never executes transactions.
python3 compare.py --amount 0.2 --max-spend 0.1 --min-rate 2000 --output report.json
Requires Python 3.11+ and Foundry 1.8.3. See the tested quick start. Archive access failures are errors, never replaced with simulated results.
02 / INSIDE THE TIME MACHINE
Rewind. Branch. Inspect.
Explore alternatives with controlled local execution. The starting state is historical; the experiments are new. This does not replay a market or predict its future.
Take another path
Restore one snapshot for each action. Check balances, allowance and pool observations before executing.
Bring back the evidence
Compare receipts, gas and exact balance changes. See which constraints pass, and what makes an action a no-go.
THE EXPERIMENT ARCHIVE
Explore earlier experiments.
Inspect synthetic stress tests or a real Uniswap experiment at pinned Ethereum state. The fork executes supplied decisions; it does not replay history, predict prices or establish profit.
Open the v0.1 report explorer Synthetic stress tests & paired EVM reports
Loading a recorded offline result…
Hold baseline—Final model equity
Circuit-breaker policy—Final model equity
Candidate − baseline—Model quote units, not USD
Hold baselineCircuit-breaker policy
Observation order → Identical inputs. No future lookahead.
| Metric | Baseline | Candidate |
|---|
Source state and action receipts
Summary balances are approximate ETH; tables preserve exact wei and token units. Symbols are labels. Full receipts, addresses and token deltas are available in the JSON below.
| Branch | Step | Status | Gas used | Native balance (wei) |
|---|
Assumptions and limits
Full result JSON
Local files stay in your browser. No uploads, accounts, analytics, wallet connection or transactions. A valid hash verifies integrity, not model correctness.
BUILD THE TIME MACHINE TOGETHER
Small repositories.
Meaningful pull requests.
Start where you can help. Improve a failure case, an API boundary or a reproduction command. A negative result is a useful contribution.
engine ↗
Execution, validation and isolation
PYTHON sdk-python ↗Client and typed report access
PYTHON cli ↗Commands and developer experience
PYTHON scenarios ↗Open fixtures and source provenance
JSON / SCHEMAS website ↗Documentation and report inspection
HTML / CSS / JS coordination ↗Roadmap, evidence and contribution map
PROJECT