Important Disclaimer
This content is for educational purposes only and does not constitute financial, investment, or trading advice. Cryptocurrency arbitrage involves significant risks including potential loss of capital. Always conduct your own research and consult with qualified financial advisors before trading.
Three Ways to Trade Arbitrage
If you search for "crypto arbitrage bot," you'll find dozens of listicles recommending specific platforms. Most of those articles earn affiliate commissions on every signup they generate. That doesn't make their recommendations wrong, but it does mean you're unlikely to read an honest comparison of whether you actually need a bot at all.
There are three distinct approaches to crypto arbitrage, and the right one depends on your capital, experience, and how much time you want to spend:
- Bots: Automated software that detects and executes trades without human intervention
- Scanners: Tools that find opportunities and show them to you, but you decide whether and when to execute
- Manual trading: Checking prices across exchanges yourself and executing trades by hand
Each has real tradeoffs. This guide covers them honestly.
If you're new to arbitrage entirely, our intro guide covers the basics before you start evaluating tools.
Arbitrage Bots: What They Actually Do
An arbitrage bot connects to multiple exchange accounts via API keys, monitors price feeds in real time, and executes buy/sell orders automatically when it detects a spread that exceeds your configured threshold.
The appeal is obvious: bots don't sleep, don't hesitate, and can react in milliseconds. In a market where profitable spreads can close within seconds, speed matters.
How They Work in Practice
- You create accounts on 2+ exchanges and complete KYC
- You generate API keys (with trade permissions) on each exchange
- You connect those keys to the bot platform
- You configure parameters: minimum spread threshold, position size, which pairs to monitor, stop conditions
- The bot monitors prices 24/7 and executes when conditions are met
Most bots require you to pre-fund accounts on every exchange they connect to. The bot can place orders but typically cannot transfer funds between exchanges for you. This means you need capital sitting on multiple platforms simultaneously.
Types of Bots Available
Exchange-native bots (free): Binance, OKX, and Pionex offer built-in bots. These are free because the exchange earns from your trading fees. The catch: they only work within that single exchange. Binance and OKX bots focus on funding rate (delta-neutral) arbitrage - going long spot and short futures on the same asset to capture funding payments. Pionex offers similar strategies plus grid bots.
Third-party bot platforms (subscription): Cryptohopper, Bitsgap, 3Commas, and HaasOnline connect to multiple exchanges via API. This enables cross-exchange arbitrage (buying on one exchange, selling on another). Pricing ranges from $23/month to $140+/month depending on the platform and plan tier.
What Bots Cost (As of May 2026)
| Platform | Type | Pricing | Arbitrage Capability |
|---|---|---|---|
| Pionex | Exchange + free bots | Free (0.05% trading fee) | Funding rate arb. Single-exchange only. |
| Binance | Exchange + free bot | Free (standard trading fees) | Delta-neutral funding rate bot. Single-exchange. |
| OKX | Exchange + free bot | Free (standard trading fees) | Smart Coin arb bot. Single-exchange. |
| Bitsgap | 3rd-party platform | $23 / $55 / $119 per month | Cross-exchange arb bot. 17+ exchanges. 7-day trial. |
| 3Commas | 3rd-party platform | $20 / $50 / $140 per month | DCA, Grid, Signal bots. Up to 15 exchanges on Pro. |
| Cryptohopper | 3rd-party platform | Free / $24 / $58 / $108 per month | Cross-exchange arb only on Hero plan ($108/mo). |
| HaasOnline | Advanced platform | Subscription (varies) | Spatial, triangular, statistical arb. Custom scripting. |
Information
Why are exchange-native bots free? Because the exchange earns revenue from your trading fees. Every bot trade generates commissions for the platform. The bot is a feature that keeps you trading on their exchange, not a standalone product.
The Real Risks of Bots
The bot listicles usually list "technical failure" as a risk and move on. The reality is more specific:
Misconfiguration is the biggest danger. A bot with a minimum spread threshold set too low will execute trades where fees exceed profit. It will do this repeatedly, automatically, 24/7, compounding losses while you sleep. Getting the parameters right requires understanding fee structures, slippage, and order book depth on every exchange you connect.
API key security. Every bot requires API keys with trade permissions on your exchange accounts. If the bot platform is compromised, attackers could execute unauthorized trades. Best practice: never grant withdrawal permissions on API keys used for bots. Keep withdrawal-capable keys separate and offline.
Bot vs bot competition. If thousands of users run the same Pionex or Cryptohopper algorithm with similar parameters, they're all competing for the same spreads. The fastest execution wins. Everyone else gets the leftover (often unprofitable) trades.
You're also competing with institutions. Professional trading firms run custom arbitrage algorithms on co-located servers with microsecond latency. Public-facing bot platforms cannot match this infrastructure. The spreads you capture are the ones institutional bots left behind.
Arbitrage Scanners: Find Opportunities, Decide Yourself
A scanner monitors prices across exchanges and presents opportunities in real time. The difference from a bot: you make the final call on whether to execute.
How Scanners Work
- The scanner connects to price feeds from multiple exchanges
- It calculates the spread between the same asset on different platforms
- It factors in trading fees (and sometimes withdrawal fees) to show net potential profit
- You see live opportunities and decide which to act on
- You execute the trade yourself on the respective exchanges
What Scanners Get Right
You learn the market. Every trade you evaluate builds pattern recognition. You start noticing which exchange pairs produce consistent spreads, what time of day opportunities appear, and when spreads look too good to be true (they usually are). A bot executing trades behind the scenes teaches you nothing.
Human judgment catches what algorithms miss. A scanner might show a 1.2% spread on an altcoin pair. A bot would execute immediately. A human trader checks the order book depth first and notices the sell side is thin enough that slippage would eat the entire spread. You skip that trade. The bot would have lost money on it.
Lower cost barrier. Many scanners are free or significantly cheaper than bot subscriptions. Some are ad-supported, some offer freemium tiers, and some (including ours) provide core functionality without paywalls.
No API keys required for scanning. You don't need to connect exchange credentials to see opportunities. This removes the security risk entirely until you decide to execute.
What Scanners Get Wrong
Speed. By the time you see an opportunity, evaluate it, and manually execute across two exchanges, the spread may have closed. This is especially true for popular large-cap pairs where bots dominate.
Availability. Scanners require active monitoring. If you're not watching, you miss opportunities. Bots don't have this limitation.
Execution friction. Manually placing orders on two different exchanges takes time. Even 30 seconds can matter.
Who Scanners Work For
Scanners are the right tool if:
- You're learning arbitrage and want to understand how spreads behave before risking capital on automation
- You have $1,000-$10,000 in capital and want to trade part-time
- You prefer to evaluate each opportunity before committing funds
- You don't want to grant API trade permissions to a third-party platform
Manual Trading: Educational but Not Practical
Manual trading means opening multiple exchange tabs, comparing prices yourself, calculating fees manually, and executing trades without any tooling assistance.
This is how most people first encounter arbitrage. You notice Bitcoin is $100 cheaper on one exchange and think "I could profit from that." You calculate fees, withdrawal costs, and transfer times by hand, and if the numbers work, you execute.
When It Makes Sense
For learning only. Manually calculating a few arbitrage opportunities teaches you more about fee structures, slippage, and timing than any tool can. Do this 10-20 times before investing in a bot or scanner.
With very small capital. If you're working with under $500, the cost of a bot subscription ($20-$140/mo) doesn't make sense. Manual execution on a few trades per week is a reasonable way to test the waters.
Why It Doesn't Scale
Profitable spreads in 2026 last seconds, not minutes. Manually scanning 5+ exchanges for price differences, running the math, and executing both legs of a trade is simply too slow. You'll see opportunities that closed before you can act on them. At any capital level above $1,000, the time cost of manual trading makes a scanner (often free) the better option.
Side-by-Side Comparison
| Bot | Scanner | Manual | |
|---|---|---|---|
| Cost | Free (exchange-native) to $140+/mo (3rd-party) | Free to low-cost | Free |
| Speed | Milliseconds to seconds | Seconds to minutes (your reaction time) | Minutes |
| Trades 24/7 | Yes | No (requires you watching) | No |
| Requires API keys | Yes (with trade permissions) | No (for scanning only) | No |
| Learning value | Low (trades happen behind the scenes) | High (you evaluate each opportunity) | Highest (but impractical) |
| Risk of misconfiguration | High (bad parameters = automated losses) | None (you decide each trade) | None |
| Minimum practical capital | $3,000-5,000+ across multiple exchanges | $500-1,000+ | Any |
| Best for | Experienced traders with capital | Most traders learning or active part-time | Learning only |
The Business Models Behind "Free" Tools
Understanding why tools are free (or not) helps you evaluate what you're actually paying for.
Exchange-native bots are free because every trade generates commissions for the exchange. Pionex charges 0.05% per trade. If their bot executes 100 trades/day for you, that's significant fee revenue for Pionex, even though the bot itself costs you nothing. The incentive alignment: the exchange profits when you trade more frequently, regardless of whether those trades are profitable for you.
Third-party bots charge subscriptions because they don't earn trading fees. Their revenue comes from your monthly payment. The incentive alignment is slightly better: they need you to stay subscribed, which means you need to perceive the bot as useful. But they still earn whether your trades are profitable or not.
Most "best arbitrage bot" articles earn affiliate commissions. When a review site recommends Cryptohopper or Bitsgap and you sign up through their link, they earn a referral fee. This doesn't make the recommendation wrong, but it explains why you'll struggle to find an article that says "you might not need a bot at all." There's no affiliate commission for that recommendation.
How to Choose
Choose a Bot If:
- You've traded arbitrage manually or with a scanner for at least a few months and understand the mechanics
- You have $5,000+ spread across pre-funded exchange accounts
- You're comfortable granting API trade permissions to a third-party platform (or using exchange-native bots)
- You understand that configuration mistakes can compound losses
- You have a strategy for monitoring bot performance regularly (don't set-and-forget)
Choose a Scanner If:
- You're learning arbitrage or have been trading for under a year
- You have $500-$5,000 in capital
- You want to understand why each trade works before committing to automation
- You trade part-time and can monitor during active trading sessions
- You prefer not to give API trade access to external platforms
Choose Manual If:
- You're researching arbitrage before investing in tools
- You want to understand fee structures, transfer times, and slippage from firsthand experience
- You have under $500 and want to test a few trades before scaling
The progression most traders follow: manual (a few weeks of learning) -> scanner (active trading with oversight) -> bot (if and when automation makes sense for your scale). Skipping straight to a bot is how most people lose money on misconfigured trades before they understand the mechanics.
Why We Built a Scanner
Transparency note: Opportuna is a scanner, not a bot. We have an obvious bias here, and you should factor that in when evaluating what follows.
We chose to build a scanner because we believe the biggest problem for most arbitrage traders isn't execution speed - it's finding real opportunities in a market flooded with noise. A scanner that shows you verified, fee-adjusted opportunities across established exchanges gives you the information advantage without taking away your decision-making.
A bot that executes a bad trade for you isn't helping. A scanner that shows you an opportunity and lets you decide whether the order book depth, fee math, and transfer timing make it viable - that's a tool that makes you a better trader over time.
If you reach a point where you've traded enough to confidently define your parameters, and you have the capital to justify automation, a bot becomes a reasonable next step. We'd rather you outgrow our tool because you've learned enough to automate profitably than stay dependent on a scanner because you never learned to evaluate trades yourself.
For more on choosing the right exchanges for your setup, see our exchange comparison. For understanding whether the margins are realistic, read is arbitrage actually profitable?.
Frequently Asked Questions
Do crypto arbitrage bots actually work?
They can, when properly configured and run on accounts with sufficient capital and liquidity. The failure mode isn't that bots don't work - it's that misconfigured bots execute bad trades automatically. Common problems: thresholds set too low, insufficient liquidity checks, and ignoring fee calculations. Always test with small amounts or paper trading before deploying real capital.
How much money do I need for a crypto arbitrage bot?
Practically, $3,000-5,000 minimum spread across 2-3 pre-funded exchange accounts. Exchange-native bots (Pionex, Binance) allow starting lower, but the profit on thin margins with small capital is minimal after fees. Third-party bots add $20-$140/month in subscription costs on top of the trading capital you need.
Are crypto arbitrage bots safe?
The bots themselves aren't the primary risk - it's the API key permissions you grant. Never give withdrawal permissions to bot API keys. Use IP whitelisting where supported. The bigger safety concern is configuration: a misconfigured bot can systematically execute losing trades. Monitor performance daily during the first month.
What's the difference between a bot and a scanner?
A bot finds opportunities and executes trades automatically. A scanner finds opportunities and shows them to you - you decide and execute manually. Bots are faster but remove human judgment. Scanners are slower but let you evaluate each trade before committing capital.
Can I start with a scanner and switch to a bot later?
Yes, and this is the approach we'd recommend. Use a scanner to learn how spreads behave, which pairs produce consistent opportunities, and how fees affect real-world profitability. Once you can define clear parameters for what constitutes a good trade, you have the knowledge to configure a bot effectively.
Why do most articles only recommend bots?
Most "best crypto arbitrage bot" articles earn affiliate commissions when you sign up for the platforms they recommend. There's no affiliate revenue in recommending a free scanner or manual trading. The recommendations may still be valid, but it's worth understanding the business model behind the advice.
See Live Arbitrage Opportunities
Real-time spreads across regulated exchanges. No API keys required to browse.
View Opportunities
