Trade IntelEvidence-backed trading automation with approvals before execution.
Trade Intel is a dashboard for building strategies, testing evidence, monitoring signals and approving trade tickets before they move into customer-owned broker, exchange or venue workflows. OpenClaw adds a Telegram command layer for users who want assistant-driven access.
Product workflow
Strategy evidence to approved ticket
A customer defines the rule, checks the evidence, monitors the signal and approves the staged ticket before anything moves to a connected account workflow.
Shares
7/8
positive rolling windows
Crypto
5/6
positive benchmark windows
Prediction
47
settled replay trades
Ticket status
Strategy
User-owned rule and limits are set
Evidence
Rolling tests and weak periods reviewed
Ticket
BUY 10 NVDA ready for review
Approval
Waiting before connector action
Choose the strategy
Bring a rule, use an AI-generated candidate or start from a tested template.
Check the evidence
Review rolling tests, weak periods, costs, caveats and promotion gates before use.
Monitor signals
Watch live rules, alerts and strategy conditions from the dashboard and connected data lanes.
Approve the ticket
Review the ticket, rationale and limits. Approve, reject or cancel before account action.
Automate approved trades
In supported connected-account lanes, approval lets Trade Intel place or queue that exact ticket.
OpenClaw extends the dashboard workflow for users who want Telegram control.
The core product is the Trade Intel web app: strategy cards, evidence, signal monitoring, ticket approvals, connector controls and audit. OpenClaw is an additional access layer for users who want to operate parts of that same workflow from Telegram.
Assistant-enabled
Users who enable OpenClaw can operate selected dashboard workflows from Telegram.
Tenant-scoped
Each bot token is tied to a workspace, permission set, market list, symbol list and notional limits.
Staged by default
OpenClaw can prepare and stage, while live submission remains approval and connector gated.
Cockpit evidence
The website records staged orders, connector status, risk checks, approvals and audit trail.
Optional Telegram examples
Natural commands, governed actions
risk check buy 10 NVDA at 950
OpenClaw returns a structured risk review before anything is staged.
stage buy 10 NVDA at 950 because breakout confirmed
Creates a pending approval ticket with rationale, limits and audit evidence.
approve order <id>
Moves the staged ticket to approved without bypassing connector gates.
submit order <id>
Routes the approved ticket through the available customer-owned paper or live-gated connector.
pause live trading
Disables live trading and turns on kill switches across the tenant connectors.
Show the proof before the rule is trusted.
Each market lane shows what has been tested, which assumptions were used, where the idea was weak and what should remain paper-only. The evidence is meant to make the next decision clearer, not to pretend historical simulations are live results.
Rolling windows
Test multiple historical periods instead of a single best-looking run.
Regime checks
Show weak markets, failed conditions, drawdowns and where the idea should not be used.
Realistic frictions
Include fees, spread, fill and liquidity assumptions where the lane supports them.
Promotion gates
Promote only when validation improves; keep failed variants out of live automation.
A practical workflow for turning ideas into decisions.
Trade Intel is not just a place to type a rule and run a backtest. Users can bring a rule, ask for AI-generated candidates or start from templates. They then choose, edit and own the final strategy card before it can monitor markets or create tickets.
The evidence section explains how a candidate is tested. This section shows the operating loop: strategy discovery, user ownership, live signals, ticket approval and continuous learning.
Strategy discovery
Turns user ideas, AI candidates and templates into customer-owned strategy cards.
Evidence engine
Tests candidates across windows, regimes, costs and weak periods before promotion.
Learning loop
Feeds outcomes, mistakes and rejected variants back into future filters and review notes.
Where ideas come from
User rule
The customer describes a market thesis, trigger, exit and limits.
AI candidate
AI proposes variants from public data, templates, factor screens and prior evidence.
Tested template
Start from a structured workflow, then edit it for the customer account and risk limit.
Customer-owned card
Nothing monitors or trades until the user accepts the final strategy card.
Lessons and self-healing
Post-trade review
Compare the ticket rationale, approval context, account action and outcome.
Lesson capture
Convert repeated mistakes into filters, warnings, sizing limits and review prompts.
Self-healing guardrails
Demote weak variants, tighten limits and require fresh evidence before reuse.
Incremental improvement
Keep useful lessons attached to the next strategy review and approval decision.
Model and automation layer
AI helps discover and review; approval controls trading
Customers see the evidence, caveats, claim status, ticket logic and audit trail before approving account action.
A ticket is managed until it is done or overridden.
It is not just an alert. A ticket can start from a strategy signal, TradingView webhook or Telegram command. It records what to trade, where to trade it, how large it is, why it fired, which checks passed and what Trade Intel is allowed to submit or queue after approval. The workflow then tracks the approved instruction until it is complete, cancelled, expired or overridden by the user.
What
Instrument, side, order type, size and destination account.
Why
Rule trigger, evidence window, AI critique and risk notes.
Limits
Max notional, stop/exit rule, market eligibility and daily controls.
After approval
Supported shares and crypto connectors can place or queue the approved trade, then keep the ticket visible until filled, cancelled, expired or overridden by the user. Prediction-market and CFD tickets require extra venue, account and jurisdiction gates before any live routing.
Submit or queue
The approved ticket is sent only to the configured customer-owned account workflow.
Track response
Accepted, rejected, queued, submitted and cancelled states stay attached to the ticket.
Watch until complete
The workflow monitors the approved instruction until it is filled, expired, cancelled or otherwise closed.
User override
The user can reject, cancel or override the workflow where the connector and venue allow it.
Current ticket
BUY 10 NVDA
Market: US shares
Status: staged from Telegram
Destination: connected broker workflow
Telegram command
“risk check buy 10 NVDA at 950” is received from the paired OpenClaw bot.
Risk gate passed
Symbol, market, notional and daily limits are checked.
Ticket built
BUY 10 NVDA, market order, rationale and risk notes.
Approved
User approves the staged order from Telegram or the cockpit.
Submitted
Approved ticket is submitted or queued through the supported connector.
Paused if needed
“pause live trading” turns on kill switches across connectors.
Bring your own data, models and trading accounts.
Connect model keys, market-data keys and supported broker or venue accounts where your plan allows it. Trade Intel keeps permissions scoped and shows exactly which approved tickets can be submitted.
Shares and crypto can automate approved tickets through supported connected-account workflows. Prediction-market and CFD tickets remain gated unless the account, venue, plan and local rules allow live routing.
Alpaca
Broker workflow
Interactive Brokers
Broker workflow
TradeStation
Broker workflow
Coinbase
Crypto venue
Kraken
Crypto venue
Binance
Crypto venue
Kalshi
Prediction venue
Polymarket
Prediction venue
OpenRouter
Model routing
Google Gemini
Model provider
Anthropic
Model provider
Polygon.io
Market data
Third-party names are used for compatibility identification only. Trade Intel is not endorsed by or affiliated with those providers unless expressly stated.
Broker accounts
Crypto venues
Prediction-market venues
AI and market data
Bot command layer
Client-owned authority
You own the strategy, account access, approvals and final decision.
Approval-gated automation
Signals create tickets. Supported account workflows can trade only after you approve the ticket.
Scoped keys and limits
Keys are scoped, masked after save and separated from research model keys.
Strategy card
Your rule, market, trigger, size, exit and limits.
Evidence check
Backtests, caveats and weak periods before use.
Signal watch
Live monitoring creates tickets when the rule fires.
Ticket package
Side, instrument, size, order type, rationale, risk checks and destination.
Approval gate
Approve, reject or cancel before anything is sent to an account workflow.
Trade automation
Where the connector, plan and permissions support it, Trade Intel submits or queues the approved ticket to the user-owned account.
Assistant access
OpenClaw can operate the same workflow by Telegram command for users who enable it.
Audit trail
Every signal, ticket and decision is recorded.
Watch the client workflow end to end.
The walkthrough leads with the core dashboard workflow: strategy setup, evidence review, signal monitoring, staged approval, connector handoff and audit. OpenClaw appears as an optional command layer near the end.
Simple product boundary.
Trade Intel can automate approved tickets in supported connected-account lanes. It does not trade without approval, take custody, issue products, act as your broker or make discretionary decisions for you.