💳 Credits Usage Guide

Learn how DeepDocs consumes credits so you can plan scans confidently and keep documentation up to date.

📊 Credit Consumption During a Scan

Credit consumption is based on number of documentation files DeepDocs has to analyse to determine if it is outdated.

A typical workflow looks like this:

First Scan: Baseline Build — After you pick a branch, source context, and target docs, DeepDocs inspects every documentation file to construct the code-to-doc map. Credits equal the total doc count (e.g., 100 docs → 100 credits) because the entire documentation surface must be analyzed once. Note that the number of detected outdated docs will be lesser than or equal to the number of analyed files
Subsequent Scans: Targeted Updates — New commits reuse the cached graph, so DeepDocs only re-scans documentation linked to the changed code. The number of credits consumed in a subsequent scan will be lesser than the first scan.

Example — A repository with 100 docs spends 100 credits on the first scan, a later commit touching one mapped doc spends 1 credit, and another commit that impacts three docs spends just 3 credits.

✅ Best Practices

Configure DeepDocs on a branch where PR review is required instead of pushing directly to main or staging; credits are best spent once code changes are stable.
Approve and merge DeepDocs PRs regularly so documentation stays current and future scans have less to analyze; merging a DeepDocs branch does not trigger another scan.
Keep the source code context narrow to reduce scan time and credit usage
Avoid deleting the scan cache unless major repository changes make prior mappings unreliable.
Monitor the dashboard after each run to keep an eye on credit usage and catch anomalies quickly.

🛠 When a Reset Is Unavoidable

Major documentation reorganizations may require clearing the cache; plan for a higher one-time credit spend when you do.
Outside of rare restructures, leave the cache intact so DeepDocs can keep scans precise and cost-effective.