Guide
When to Use ATF
Not every bot needs every feature. This guide helps you decide when to start, which mode to use, and when upgrading makes sense.
Mock vs Real Execution
ATF supports two execution modes. Choosing the right one depends on your stage.
| Mode | What Happens | When to Use | Cost |
|---|---|---|---|
| Demo / Mock | ATF evaluates your intent and returns a receipt, but no transaction is sent to Solana. | First integration. Testing policy rules. CI pipelines. Validating receipt format. | Free - no signup required |
| Real Execution | ATF evaluates, and if allowed, the transaction proceeds to Solana. Full enforcement with on-chain settlement. | Production bots executing real capital. After you have validated policy rules in mock mode. | Pro tier (manual approval in beta) |
Start with mock mode. Validate your policy configuration and receipt handling. Move to real execution when your bot is proven and you need on-chain settlement.
When Free Is Enough
Free tier includes 100 protect calls per day, mock execution, and full receipt generation. This is sufficient for:
- Evaluating ATF for your bot - test policy enforcement, verify receipt format, check integration patterns.
- Low-frequency bots - bots that execute fewer than 100 trades per day and operate within default spend caps (25 SOL per tx).
- CI/CD testing - integrate protect calls into your test suite to verify policy compliance before deployment.
- Agent framework testing - validate that your agent correctly handles ALLOW/DENY responses and processes receipts.
- Demos and proof-of-concept - show stakeholders how enforcement works without committing to a paid plan.
If your bot never needs to exceed 100 calls/day and 25 SOL per transaction, Free may be all you need.
When Pro Becomes Useful
You will know Pro is useful when your bot starts hitting limits. The deny receipts tell you exactly when that happens.
- Rate limit exceeded - your bot makes more than 100 protect calls per day.
- Spend cap exceeded - your trades exceed the 25 SOL per-tx cap on Free.
- Real execution needed - you want ATF-gated real settlement on Solana, not just mock evaluation.
- Higher-frequency trading - arbitrage bots, market makers, or agents that execute continuously.
- Custom policy configuration - tighter or more specific rules than the defaults allow.
During public beta, Pro requires manual approval. Request an upgrade when you are ready.
What Kinds of Bots Benefit First
ATF is designed for any bot that executes transactions on Solana, but some use cases benefit more immediately than others.
| Bot Type | Why ATF Helps | Start With |
|---|---|---|
| Arbitrage bots | High-frequency, high-risk. Spend caps prevent a bad arb from draining the wallet in one trade. | Free → Pro when hitting caps |
| DCA / scheduling bots | Predictable execution. Policy ensures each scheduled trade stays within bounds and receipts provide an audit trail. | Free (usually sufficient) |
| AI agent frameworks | Agents make autonomous decisions. ATF adds a deterministic constraint layer before execution. | Free + OpenClaw plugin |
| Treasury / fund automation | Larger position sizes. Protocol allowlists and spend caps enforce fiduciary constraints. | Pro or Enterprise |
| Liquidation bots | Time-sensitive. TTL-bound permits ensure stale decisions do not execute. Receipts prove every action. | Pro (real execution needed) |
Quick Decision Flow
- Want to see how ATF works? → Use the web sandbox (no signup)
- Integrating into your bot? → Start with Getting Started (Free tier, mock mode)
- Hitting limits on Free? → Your deny receipts will tell you. Then request Pro
- Need real execution on Solana? → Pro tier with manual approval during beta
- Enterprise compliance needs? → Contact us