Q&A with PM Wisdom. PolyOrderbooks in the hot seat: what we archive, why one-second depth matters for research, how the MCP server fits AI workflows, and where simulated backtests stop being reliable. First published as a featured-listing interview on PM Wisdom.
PolyOrderbooks archives order books, prices and liquidity metrics for Polymarket crypto markets. Researchers can query the data directly or use its AI agent and Strategy Builder to test a strategy without writing a backtesting engine. This Q&A covers those workflows, plan limits and what simulated results can miss. Product overview, Backtesting.
What does PolyOrderbooks do?
We keep a history of Polymarket crypto markets that people can query after the market has moved on. That includes prices, the buy and sell orders available at different prices, and liquidity metrics such as volume and spread. You can compare those datasets on one shared one-second timeline or use the archive to run a simulated backtest. PolyOrderbooks is an independent product and isn't affiliated with Polymarket. Website, Backtesting.
Who is it for?
Researchers studying how these markets behave, quants testing strategies, and developers building tools around historical data. Someone might want to examine a price move, check how much liquidity was available, or bring the archive into a research notebook. The API provides the underlying data. The AI agent and Strategy Builder also let people test rules without building that workflow in code. API overview, Backtesting.
Why do order books matter?
A price chart doesn't tell you how much you could buy or sell at that price. The order book shows the available size at each level. A larger order may need to take several levels, which changes its average execution price. Looking at that depth helps you spot assumptions about fills that a simple price-based backtest would miss. Working with order books.
Why one-second resolution?
It lets you examine changes inside a minute. Books, prices and liquidity metrics can all be queried on the same one-second timeline, so you can compare a price move with the depth and spread around it. For longer-range research, coarser intervals are available too. The useful resolution depends on the question you're asking. Historical data overview, Query resolutions and plans.
How do people work with the data?
The REST API works with any language that can make HTTP requests. Our official Python SDK wraps those endpoints for scripts and notebooks. A typical workflow starts by finding a market in the catalog, choosing a time window, then fetching its history. Longer result sets are paginated so you can retrieve them in batches. Python and API quickstart.
What does MCP add?
Our official MCP server makes the archive accessible from tools such as Cursor and Claude Desktop. You can use those clients to find markets and request historical books, prices and metrics without writing each API call yourself. The package is @polyorderbooks/mcp-server, and it uses your PolyOrderbooks API key. The same plan limits apply. MCP server.
What are the plans and limits?
Starter is free and includes one-second order books, prices and liquidity metrics. The current self-serve paid plans are sold by history window; the documentation calls these Pro plans. All listed data plans allow one-second queries.
| Data plan | Monthly price | API history | Requests/minute | Requests/day |
|---|---|---|---|---|
| Starter | Free | 3 days | 60 | 1,000 |
| 30-day / Pro | $19 | 30 days | 300 | 50,000 |
| 60-day / Pro | $29 | 60 days | 300 | 50,000 |
| 90-day / Pro | $39 | 90 days | 300 | 50,000 |
| 120-day / Pro | $49 | 120 days | 300 | 50,000 |
| Enterprise | Custom | Custom | Custom | Custom |
Prices and limits checked on 11 September 2026. Current pricing, Plan documentation.
Fast adds $19/month for 1,000 requests/minute and 120,000/day. Heavy adds $39/month for 2,500/minute and 300,000/day. Speed packages.
API caps are per minute and per day. Monthly quotas apply to backtests:
| Backtesting access | Monthly allowance | Backtesting history |
|---|---|---|
| Starter | 1 AI + 3 strategy backtests | 1 day |
| Paid data window, without Backtest AI | 1 AI + 3 strategy backtests | Selected data window |
| Backtest AI add-on, +$19/month | 500 AI + unlimited strategy backtests | Selected data window |
| Backtest AI standalone, $29/month | 500 AI + unlimited strategy backtests | 30 days |
Starter's one-day backtesting window is separate from its three-day API window. Backtesting allowances.
Do I need to write code?
No. You can describe a crypto up/down strategy to Backtest AI in plain English, or enter rules in the structured Strategy Builder. Both run simulated backtests against the same one-second archive used by the API and SDK.
The simulation uses historical order book depth to model partial fills and slippage. The report includes net P&L, win rate, profit factor, maximum drawdown, an equity curve and a trade log with exit reasons. Each run includes an AI-written analysis. You can export trades as CSV and download the report as JSON. Backtest AI overview.
What should I be careful about?
Our coverage is Polymarket crypto markets. Historical time buckets can repeat the last captured value, so a timestamp doesn't necessarily mean a new book update arrived at that instant. The archive isn't a record of every individual order-book event. Historical data semantics.
Execution results are estimates. Available depth matters, and a simulation still needs explicit assumptions about how orders fill, including where an order sits in the queue. Backtesting methodology.
Historical results don't establish what a live strategy will earn.
How do I get started?
Create an account and an API key, then pick one market and a short time window. Try a query through REST, Python or MCP and inspect the returned data before building a larger workflow. Password sign-ups need email verification; Google sign-ups are verified automatically. The quickstart walks through account setup and a first historical query. Quickstart.
Prepared by PM Wisdom. Reviewed by PolyOrderbooks on 10 September 2026. See the PolyOrderbooks featured listing and explainer.