10x vs Grepr
Grepr sends your stack samples and summaries of your logs.
10x keeps every line searchable, in your stack or S3, at a flat price.
Splunk
| 10x | Grepr | |
|---|---|---|
| Where logs are processed | In your network, before logs ship | Grepr’s AWS, both deployment models |
| What Splunk receives | Full, searchable originals, compacted in place · 10x Splunk app Learn how | Samples and summaries, not every line |
| Where raw logs are kept | Splunk, or your S3 | A second store Grepr hosts and meters |
| Getting logs back | From your S3: selected lines, in seconds; re-ingest optional | Backfill from Grepr’s store, billed as new ingest |
| Who writes the config | Your AI model, in your account | Grepr’s models, in Grepr’s cloud |
| Pricing | Per node, published; volume spikes cost nothing extra | Grows with volume; no published rate card |
Data Control
Logs leave your network; Grepr stores a full copy and meters queries against it.
Splunksamples + summariesLogs are reduced in-network and reach only the platform you already pay for and your S3.
Splunkevery line kept, compacted · smaller billFrequently Asked Questions
Quick answers on Log10x vs Grepr
What is the difference between Log10x and Grepr?
10x cuts log costs in place on the destination you already pay for, keeping every line queryable there or in your own S3, and it exposes a per-pattern cost series in your own Prometheus. Grepr clusters lines and forwards samples plus one summary per pattern per window, with changing values swapped for placeholders like <any> and <number>, while the full raw is kept in a separate Iceberg or Parquet lake that, by default, runs in Grepr's own cloud with metered query volume.
Is Log10x's cost reduction lossless?
It keeps every line, but losslessness is per-destination. Compacting is a lossless re-encode on Splunk, self-hosted Elasticsearch or OpenSearch, and ClickHouse, and a no-op on Datadog and CloudWatch. Offload to your own S3 is the universal lever and keeps every line on every destination. Sampling and dropping are a last resort, not the default.
Does Grepr keep my logs in my own network?
By default, no. Grepr's processing engine runs in Grepr's own AWS VPCs and its pricing meters data ingestion volume (SaaS), so raw telemetry transits Grepr's cloud; bring-your-own-bucket covers the lake storage, not the processing path. Grepr's own docs say it runs in AWS for both deployment models, and no self-hosted install is publicly documented. 10x classifies and acts inside your own network and offloads to your own S3, so a fully in-network deployment is possible.
Does Grepr reduce metrics, or only logs?
Grepr reduces metrics and traces as well as logs; it is not a logs-only tool. The difference between Log10x and Grepr is not which signals they cover, it is where the reduced data ends up: 10x keeps a per-pattern cost series in your own Prometheus and cuts cost in place, while Grepr summarizes into its own backend and lake.
Can I query a per-pattern cost series in my own TSDB?
Yes. 10x exposes every message type as a durable per-pattern time series in your own Prometheus, so you can query volume, cost, and trend per pattern in the TSDB you already run. Grepr's summary rows are keyed by an internal patternId inside its backend.
How much can Log10x cut a log bill?
On a large self-hosted Elasticsearch cluster where _source is about 95 percent of storage, re-encoding the structured logs in place models to a conservative 20 percent index-storage reduction, with every line still queryable in the same cluster. On Datadog, where compact is a no-op, the saving comes from offload and tier-down instead. These are modeled figures with a shown method, not a guaranteed percentage.
When is Grepr the better choice?
If a push-button recovery experience matters more than keeping data in place, Grepr's backfill replays matched raw from its lake in your existing Datadog or Splunk query syntax and can be triggered by an alert. 10x makes the opposite trade: the data never leaves the store it is already in, and offloaded lines are read back from your own S3 in place, in seconds, with re-ingest optional.
What does Log10x not claim versus Grepr?
Compacting is lossless only on Splunk, self-hosted Elasticsearch or OpenSearch, and ClickHouse, and a no-op on Datadog and CloudWatch; offload is the universal lever. Alert-triggered backfill is Grepr's feature; 10x brings offloaded originals back from your S3 on demand. The percentages here are modeled, not guaranteed.
How is 10x priced compared to Grepr?
Grepr charges for compute plus the volume you send, and publishes no rate card. 10x is a flat fee per node, published, and does not rise with log volume.
Can the savings be verified?
With 10x you set a budget and approve the plan it proposes; savings are then measured per pattern over time, so they can be checked rather than taken on faith. Grepr auto-tunes inside its own cloud and commits to no number, stating that results depend on how noisy your logs are.
Who writes the config?
With 10x, you set a budget and any protected patterns; the 10x MCP turns that into a per-pattern action plan using your own AI agent in your environment, delivered as a GitOps pull request or straight to the cluster, and the 10x Engine enforces it. No log content goes to an outside model. Grepr tunes its filters with its own models inside its own cloud.
Is Log10x the mathematical function log10(x)?
No. Log10x is a log and observability cost-reduction company and product. It is not the logarithm log10(x), and it is not Log10.io.