NEW Pine v6 strategy research kit

Turn one trading idea into
100 tested Pine strategies.

Describe your setup in plain English. The lab builds Pine v6 variants, tests them on your own TradingView, kills the weak ones, and hands you source + evidence for every survivor.

Full source · AI-agent ready · Runs locally · Works on free TradingView · 7-day refund

No signal promises. No black box. PASS only means “promising — validate further,” never “profitable.”

251built & tested
55survived
196didn’t
0hidden
lab/dashboard.html — THE NETWORK · Mission Control
Candidates judged251
Deflated crown · t-stat4.76PASS
Killed · cause on file71×2 buries
THE NETWORK · live research log▶ time-replay
donchian-breakout-021 · PASScapitulation-reversal-008 · WEAK ⏸mean-reversion-002 · KILL
Explore the full dashboard →
The shift

Stop asking AI for one lucky script. Give it a lab.

Most tools show you the one chart that survived and hide everything that failed. This runs the whole search, saves the graveyard, and gives you a repeatable way to test your own ideas — hands-free after a ~10-minute setup.

  • 1Type your idea

    “Test a pullback after a breakout on gold, 1h.” Plain English — no Pine coding, no backtester to wire up.

  • 2The lab builds & tests it

    It writes Pine v6 variants, runs them on your real TradingView, applies the kill-gates, and logs every verdict.

  • 3You keep the source & evidence

    Survivors come with copy-paste Pine, a dossier, the metrics, and the next validation step. The whole run is yours.

  • Hundreds of strategies a night
  • Agent-readable skill + docs
  • You own every line
  • Tested on your real TradingView
  • Non-repainting signals
  • Real 251-candidate lab
  • 7-day refund

Works with the tools you already use

Best experience
Also works withAny AI agent that can run shell commands
Manual modeNo AI agent required
After-pass outputTradingView alerts · Telegram relay

All product names & logos are trademarks of their respective owners, shown to indicate compatibility. ProEA Lab is not affiliated with or endorsed by them.

What you’re buying

Not 251 strategies. The machine that makes them.

Point it at your symbol, your idea — and it builds and tests it hands-free (after a ~10-minute setup), faster than you could draft one by hand. The 251 here are just the starter demo: the same machine builds many more, all from full source you own and extend with your AI agent.

UNDER THE HOOD

A structured Pine v6 block library

18 entries × 8 filters × 9 exits, long/short — composable into 3,888 starting strategies. Many variants, strict gates, no blind optimization.

YOUR IDEA, BUILT

Describe it in plain words. The agent builds it.

No Pine coding, no backtester to wire up — you just talk to an AI agent.

THE GRAVEYARD

Every kill on file

It tests on your real TradingView and buries what doesn’t hold — cause of death logged. Most die. That’s the proof it’s honest.

Most tools sell you a strategy. This one gives your AI a quant’s spine — and the whole factory is yours to keep. See the full workflow →

Before & after

You already know the old way breaks.

If you’ve ever asked an AI for a Pine script, you know exactly how this goes.

Asking AI for a script
  • It invents functions that don’t exist
  • One backtest looks amazing
  • You can’t tell skill from luck
  • No record of what failed
  • You start over from scratch every time
Running it as a lab
  • Describe the idea once, in plain words
  • It builds structured Pine v6 variants
  • Kill-gates bury the weak ones for you
  • Survivors come with a full dossier
  • You keep the source — the graveyard teaches the next run
AI-agent ready

Agent-ready from the first folder open.

Not a locked indicator or a PDF course — a full-source workspace your AI agent can read, explain, run, and extend. Open it in Claude Code, Codex, Cursor, Gemini, or any agent that runs shell commands.

You type — in your agent

“Read SKILL.md and AGENTS.md, explain this lab to me — then run my first discovery cycle on XAUUSD 1h.”

It starts from SKILL.md · AGENTS.md · START_HERE_AI.md — the protocol, the rules, the prompts, and the full source.

  • Explain the whole system in plain English
  • Run the setup checklist
  • Build & test Pine v6 variants
  • Read the graveyard and the learnings
  • Extend entries, filters, exits, gates & recipes
  • Modify any file — you own every line
The funnel nobody publishes

We built 251. Only 55 survived.

Most products show you the one winner. This gives you the winner, the rejects, the reasons, and the machine that made them — every verdict on file, re-runnable on your machine.

251built & tested
55survived
71killed — cause on file
125honest “not yet”
55 PASS68 WEAK57 INSUFFICIENT71 KILL
PASS
promising enough to validate further — not “ready for real money.”
KILL
failed a gate; the reason is saved on file.
OOS
out-of-sample — data the strategy never trained on.
PF
profit factor — above 1.0 means gross profit beat gross loss.
Retention
how much of the edge survived outside the original test.

Every PASS, WEAK, INSUFFICIENT and KILL is in plain files. See every candidate, the live network, and the leaderboard →

Mission Control

Three views do the heavy lifting.

The dashboard has a lot more — but these are the three you’ll actually live in.

THE GAUNTLET

25155 PASS → 71 KILL

The whole run as one funnel — every candidate judged, nothing staged.

SURVIVOR DOSSIER

Click a PASS → source + evidence

Windows, real costs, frozen params, and copy-paste Pine for each survivor.

AI MISSIONS

Copy a mission → paste to your agent

Each mission carries the lab’s honesty rules, so your library keeps growing safely.

Open the full Evidence dashboard →

In the box

Not just source code. Source your AI can understand.

Everything in plain files — skills, prompts, playbooks, 15 tools, recipes, docs, the 251-candidate demo lab, and full editable Pine source. Open the folder, check it before you trust a single claim.

what ships in the box
pine-evolution-lab/
AGENTS.mdrules for Claude · Codex · Cursor · Gemini
START_HERE_AI.mdfirst prompt + first-cycle guide
PROMPTS.mdcopy-paste research missions
PLAYBOOK.mdhow to run the lab properly
EDGE-FINDER.mdidea → hypothesis → test
CHEATSHEET.md · AFTER-PASS.mdcommands · post-survivor steps
pine-evolution-lab/SKILL.md · references · 15 tools
recipes/12 one-command strategies
docs/ · auto-pilot/full docs + TradingView bridge
lab/REAL demo · 251 candidates · graveyard
  • The complete 7-stage loop — 15 dependency-free Node tools
  • Strategy builder: 18 entries × 8 filters × 9 exits, Pine v6
  • Real 251-candidate demo lab · 56 fund-style dossiers · 17 case studies
  • THE NETWORK living graph + EDGE-MAP Mission Control dashboard
  • AFTER-PASS go-live: ToS-clean TradingView alerts + Telegram relay
  • 6-step kickstart + 12-recipe “strategy in one command” pack
Fit check

Built for some traders. Honest about who it isn’t for.

It’s a research lab, not a signal feed. The right buyer will love it — the wrong buyer should not buy it.

Built for
  • AI-assisted traders — you already use Claude, Cursor, Codex, or ChatGPT to explore ideas.
  • Pine builders — you want editable source, not a locked indicator.
  • System traders — you care about evidence, rejection, and repeatable tests.
Not for
  • Signal buyers — it doesn’t send buy/sell calls.
  • Profit-promise seekers — PASS means promising, not profitable.
  • People who hate setup — you’ll run a local kit with TradingView Desktop and Node.
Get the lab

Buy the lab. Extend it with your AI agent.

One payment. A full-source workspace your agent can read and improve, the field manuals, and a real 251-candidate lab to stand on. v1.59.0.

PINE EVOLUTION LAB · CAPSTONE
◆ LAUNCH PRICE · FIRST 50
$79$149SAVE $70one-time

Launch price for the first 50 buyers — regular $149 after.

  • Open it in Claude Code, Codex, Cursor, Gemini — or run it by hand
  • Let your agent read SKILL.md · AGENTS.md · START_HERE_AI.md
  • Ask it to explain the lab before it runs anything
  • Build & test Pine v6 variants from plain-English ideas
  • Review every PASS · WEAK · INSUFFICIENT · KILL + open dossiers
  • Extend the blocks, recipes, prompts & tools — every file is yours

Educational research software. Not financial advice. No profit promise. Automation & TradingView-ToS details in the FAQ →

FAQ

Fair questions, straight answers.

What do I actually need to run it?+
TradingView Desktop (the free plan works) launched with a debug flag, Node 16+, and the Auto-Pilot bridge — bundled in the box (MIT) with a one-shot installer for Windows and macOS/Linux (~1 minute). An AI agent is recommended, not required.
Does it optimize parameters to find the “best” settings?+
Deliberately no. Re-optimizing on the engine is how you overfit. The lab's way is coarse, honest mutations — ≤2 parameters per child, judged on plateaus not spikes, with a trials counter that makes data-snooping visible in every verdict.
Why might a candidate come back INSUFFICIENT?+
Because your chart only loads so much history. The lab sizes windows to what actually loaded and refuses to pretend otherwise — windows the data can't answer print as ? instead of fabricating a number.
Do I need the companion products?+
No. Pine Strategy Foundation (idea → genome) and Pine Quant Studio (deep statistics for PASSes) make the loop stronger, but the lab runs complete on its own.
Is driving TradingView like this allowed?+
Measurement automates your own desktop app on your own account, for research — a TradingView-ToS gray area; see the automation notice in LICENSE.md, and every stage can be run by hand. Going live uses TradingView's official alerts — fully within their terms.
Will it make me money?+
Honest answer: nobody can promise you that, and we won't. The lab gives you the best-evidenced hypothesis it found, the full ledger of what died, and the explicit next gates before real size. It's educational decision-support software — not financial advice.
Start my first lab cycle — $149 $79