Getting started

The Open Backtest API lets you fetch market candles, create backtest sessions, push trades and read computed statistics — everything you see in the app, scriptable. Sessions created through the API show up in your workspace like any other, tagged api/sdk.

Authentication

The API is part of the Pro plan. Create a key in Settings → API in the app — the full key (obk_…) is shown exactly once. Pass it on every request in the x-api-key header:

curl https://api.openbacktest.fefe-du-973.fr/api/v1/me \
  -H "x-api-key: obk_your_key_here"

Base URL for this deployment: https://api.openbacktest.fefe-du-973.fr/api/v1. It is resolved from the deployment environment, so local and production documentation stay aligned.

Errors and rate limits

Every non-2xx response uses one envelope:

{ "error": { "code": "NOT_FOUND", "message": "Session not found" } }

Codes: UNAUTHORIZED, PRO_REQUIRED, RATE_LIMITED, NOT_FOUND, UNKNOWN_SYMBOL, SESSION_NOT_ACTIVE, VALIDATION_ERROR, MARKET_DATA_UNAVAILABLE, INTERNAL.

Keys are limited to 1000 requests/hour (HTTP 429 with a Retry-After header beyond that). Candle requests draw from an additional per-user budget shared with the app — one upstream quota per account.

A complete run in curl

BASE=https://api.openbacktest.fefe-du-973.fr/api/v1
KEY=obk_your_key_here

# 1. Who am I?
curl -s $BASE/me -H "x-api-key: $KEY"

# 2. Create a project
PROJECT=$(curl -s -X POST $BASE/projects -H "x-api-key: $KEY" \
  -H "Content-Type: application/json" \
  -d '{"name":"My strategy","symbol":"BTCUSDT","timeframe":"1h"}')
PROJECT_ID=$(echo "$PROJECT" | jq -r .project.id)

# 3. Open a session (fees in percent per side)
SESSION_PAYLOAD=$(jq -nc \
  --arg projectId "$PROJECT_ID" \
  --arg name "Run 1" \
  '{projectId:$projectId,name:$name,initialBalance:10000,
    feesConfig:{entryFee:0.1,exitFee:0.1,type:"percent"}}')
SESSION=$(curl -s -X POST $BASE/sessions -H "x-api-key: $KEY" \
  -H "Content-Type: application/json" \
  -d "$SESSION_PAYLOAD")
SID=$(echo $SESSION | jq -r .session.id)

# 4. Push closed trades (batches of up to 500)
curl -s -X POST $BASE/sessions/$SID/trades -H "x-api-key: $KEY" \
  -H "Content-Type: application/json" \
  -d '{"trades":[{"clientId":"strategy-run-42-trade-1",
       "side":"long","qty":0.1,"entryPrice":60000,
       "entryTime":1751000000000,"exitPrice":61000,"exitTime":1751003600000}]}'

# 5. Finish and read the stats
curl -s -X POST $BASE/sessions/$SID/finish -H "x-api-key: $KEY" | jq .stats

All dates are millisecond epochs. Trades are ingested closed (entry + exit in one record) — metrics (fees, P&L, R-multiple) are computed server-side and frozen at ingest.

Prefer Python? The SDK wraps this whole lifecycle in a few lines, with candles as pandas DataFrames.

Authentication

Errors and rate limits

A complete run in curl

On this page