# Maple vs Axiom

Axiom and Maple both take OpenTelemetry in, bill on usage, and charge nothing per seat. The difference is what sits on top of the store. Axiom gives you APL and a dashboard builder. Maple ships the surfaces built, including the service map, session replay and web analytics, and you can self-host all of it. Below: where the two diverge, what Axiom leaves you to build, where Axiom is ahead, and what the same month costs on each.

## Where Maple and Axiom differ

Every cell is a sentence. The gutter says who has the edge on that row; rows where both tools do the same thing are collapsed into one line at the end.

### Where they differ

| Topic | Maple | Axiom | Edge |
| --- | --- | --- | --- |
| Pricing model | $39 a month with 100 GB per signal included, then $0.30 per GB. Browser sessions are the only other meter. Queries and storage are not separate line items. | $25 a month, then three meters: data loading from $0.12 per GB, query compute from $0.20 per GB-hour, storage $0.03 per GB. The free allowance is 1 TB, 100 GB-hours and 100 GB. Single sign-on costs $100 a month, RBAC and the audit log $50 each. [1] | Maple |
| Error workflow | Exceptions become issues with an owner, a state, a severity and comments. A resolved issue that fires again reopens. | No issue object. Exceptions stay events: you query them in APL, group by message, and have a monitor notify per group. [4] | Maple |
| Frontend and mobile | Browser session replay with the trace behind every click, web analytics from the browser SDK, and a Swift SDK for iOS. | Web Vitals from the browser and a zero-config app for Vercel. No session replay, no mobile SDK. [9] | Maple |
| Source code | On GitHub under FSL-1.1, converting to Apache 2.0 two years after each release. | Closed. The SDKs, Collector configurations and the Vercel app are public. The platform is not. | Maple |
| Self-hosting | Yes. The whole platform on your own infrastructure, your own ClickHouse, retention you set. | No. Axiom runs the edge deployments, in US East and EU Central. Enterprise can point storage at its own S3 bucket. [3] | Maple |
| Local development | One binary with an embedded ClickHouse. `maple start` on a laptop, no containers, no account. | Cloud only. Local work still means a dataset in a real organization and a token in your environment. | Maple |
| Query languages | One visual builder across every signal. When you want the query itself it is ClickHouse SQL, through the API or the MCP server. | APL for events, MPL for metrics. Over a hundred functions, plus a skill that translates Splunk's SPL into APL. [5] | Even |
| AI and agents | A hosted MCP server over the product's own queries, scoped to the org. Investigations run the diagnosis and write up what broke. | An MCP server, six agent skills (SRE, metrics, dashboards, alerting, cost control, SPL translation) and Spotlight, which diffs two sets of events. Query cost limits per token stop an agent running up the bill. [6] | Even |
| Cost controls | Overage caps per signal and usage alerts in the billing settings. At a cap the ingest gateway rejects that signal until the billing cycle resets. | Spend alerts and hard caps in the console. Past the limit, ingestion pauses instead of billing as overage. [1] | Even |

### Where Axiom is ahead

| Topic | Maple | Axiom | Edge |
| --- | --- | --- | --- |
| Price at high volume | $0.30 per GB at every volume. Lower rates exist on Enterprise but are not published. | The ingest rate falls as volume grows, with no negotiation: $0.12 per GB, $0.085 above 500 TB, $0.06 above 5 PB. Pre-purchased credits on a 12- to 36-month term take a further discount. [2] | Axiom |
| Retention | 30 days on the cloud plan at the flat rate. Longer on Enterprise, and your call when you self-host. | Configurable per dataset and billed as storage: $0.03 per compressed GB a month for the first 60 days, $0.0163 to 120 days and $0.0052 after that, with 100 GB free. [2] | Axiom |
| Integration catalog | OpenTelemetry receivers plus a catalog of connected providers: Cloudflare, Prometheus, PlanetScale, WarpStream, Hazel and GitHub. Kubernetes and Docker are agent installs. | Over forty apps, including Vector, CloudWatch, Lambda, Heroku and CI/CD. A Splunk app federates search across both tools. [7] | Axiom |
| Compliance attestations | SOC 2 and ISO 27001 certification in progress. Maple Cloud runs in the US or the EU (Frankfurt) at the same price, and self-hosting keeps the data in your own audited environment. | SOC 2 Type II, ISO 27001, GDPR, a self-serve HIPAA BAA, and a choice of US or EU residency. [8] | Axiom |

**Both have:** Distributed tracing · Log management · Metrics and dashboards · Alerting · Kubernetes monitoring · Public API · MCP server for agents · OpenTelemetry ingest

## What Axiom leaves you to build

Axiom stores events and hands you a dashboard builder. Every view is a panel you write in APL. Maple ships these four built, and none of them is a chart you could assemble out of panels.

- **Service map**: Every service, queue and database, drawn from the spans themselves. There is no topology file to maintain. Edges carry throughput, error rate and latency, and clicking one opens the traces behind it. On Axiom: no dependency graph. You aggregate caller and callee pairs in APL and read the topology as a table.
- **Browser sessions and replay**: Watch the session back with its DOM, console, network and errors, and open the backend trace behind any click. Masking happens in the browser, before anything leaves the page. On Axiom: nothing to build on. Axiom stores browser events if you send them, but there is no capture and no replay.
- **Web analytics**: Visitors, page views, referrers and countries, from the same browser SDK that is already recording sessions. One script, and every number opens into the sessions behind it. On Axiom: a dashboard you assemble from APL over your own events, with no path from a number to the session that produced it.
- **Kubernetes console**: Workloads and restarts with live CPU and memory per pod, joined to the traces and logs of the service running inside them. On Axiom: a Kubernetes app that ships the data, and dashboards you build over it.

## The same month, billed twice

A reference workload priced on both at list price. Change the sliders underneath for your own numbers.

**Reference workload:** Log volume 100 GB/mo · Trace volume 100 GB/mo · Metric volume 100 GB/mo · Query compute 250 GB-hrs/mo

| Maple | Amount |
| --- | --- |
| Startup plan: 100 GB per signal included | $39 |
| Team seats: No per-seat fees | Free |
| **Total** | **$39** |

| Axiom | Amount |
| --- | --- |
| Platform fee: $25/mo Axiom Cloud | $25 |
| Data loading: 300 GB (1 TB free) | Free |
| Query compute: 250 GB-hrs (100 free) | $30 |
| Storage: 15 GB compressed (100 GB free) | Free |
| **Total** | **$55** |

$16 less per month on Maple

List prices checked September 2026. Sources at the foot of the page. Maple pricing based on the Startup plan ($39/mo with 100 GB included per signal (logs, traces, metrics), then $0.30/GB, billed per signal), metered on uncompressed (decoded OTLP) bytes. Axiom is modeled on Axiom Cloud's rate card: $25 a month, data loading at $0.12/GB beyond 1 TB free, query compute at $0.20/GB-hour beyond 100 GB-hours free, and storage at $0.03/GB a month beyond 100 GB at Axiom's documented 95% compression and one month of retention. Pre-purchase credit discounts and the SSO, Directory Sync, RBAC and audit log add-ons are not modeled. Maple has no query or storage meter, so its estimate is ingest only.

Interactive calculator: https://maple.dev/compare/axiom#calculator

## Moving from Axiom

Telemetry moves with one exporter change. Dashboards and alert rules don't cross vendors, so plan those as the real work.

1. **Keep the Collector, add an exporter** Both platforms ingest standard OTLP, so SDKs, auto-instrumentation and semantic conventions carry over unchanged. Add Maple next to the Axiom exporter and ship to both until the alerts you care about fire from Maple.
2. **Move the dashboards and monitors with one prompt** Dashboards and monitors do not travel over OTLP, and neither does APL. Both platforms expose an MCP server: Axiom's reads what you have, Maple's creates it. Connect an agent to both and ask it to port everything, translating the queries as it goes.

```yaml
# otel-collector.yaml. Add Maple as a second exporter, run both, then drop the first.
exporters:
  otlphttp/axiom:
    compression: zstd
    endpoint: https://api.axiom.co
    headers:
      authorization: Bearer ${env:AXIOM_API_TOKEN}
      x-axiom-dataset: ${env:AXIOM_DATASET}
  otlphttp/maple:
    endpoint: https://ingest.maple.dev
    headers:
      x-maple-ingest-key: ${env:MAPLE_INGEST_KEY}

service:
  pipelines:
    traces:
      exporters: [otlphttp/axiom, otlphttp/maple]
    logs:
      exporters: [otlphttp/axiom, otlphttp/maple]
    metrics:
      exporters: [otlphttp/axiom, otlphttp/maple]
```

## Common questions

### Is Axiom open source? Can I self-host it?

No to both. Axiom publishes its SDKs, Collector configurations and the Vercel app, but the platform is closed and runs only in Axiom's own edge deployments, today US East and EU Central. Enterprise customers can keep object storage in their own S3 bucket, which is data residency rather than self-hosting. Maple's source is on GitHub under FSL-1.1, converts to Apache 2.0 two years after each release, and runs on your own infrastructure with your own ClickHouse.

### How is Maple different from Axiom?

Axiom is an event store first. Everything lands in a dataset, you query it in APL, and a monitor runs that query on a schedule. Maple is the product on top of that storage. Exceptions become issues with an owner and a state, browser sessions replay with the trace behind each click, agent sessions are a signal of their own, and one binary runs the whole thing on a laptop. Axiom is ahead on price at volume, retention economics, integration breadth and compliance.

### Which is cheaper, Axiom or Maple?

At the reference workload Maple is, because Axiom's platform fee and query-compute meter start charging before Maple's overage does. Above a terabyte a month Axiom's ingest rate is the lower one and keeps falling with volume, and its storage pricing makes long retention cheap. The receipts above run both published rate cards on the same month, and the calculator lets you change the volumes.

### Does Axiom have error tracking or session replay?

Not as products. Exceptions are events you query in APL and group inside a monitor, so nothing carries an owner, a state or a reopen. There is no session replay. Maple has both: issues with a lifecycle, and browser replay masked in the page before anything leaves it.

### Do I have to learn APL?

On Axiom, mostly yes. APL for events, MPL for metrics, and a skill that translates SPL if you are arriving from Splunk. Maple's visual builder covers every signal, and when you want the query itself it is ClickHouse SQL, through the API or the MCP server.

## Where the Axiom claims come from

Vendor facts on this page link to the vendor's own pages and carry the month they were last checked. If one is stale, tell us.

1. [Axiom pricing](https://axiom.co/pricing), checked September 2026
2. [Axiom rate card (the machine-readable one)](https://axiom.co/pricing/llms.txt), checked September 2026
3. [Axiom: edge deployments](https://axiom.co/docs/reference/edge-deployments), checked September 2026
4. [Axiom: monitors](https://axiom.co/docs/monitor-data/monitors), checked September 2026
5. [Axiom: APL](https://axiom.co/docs/apl/introduction), checked September 2026
6. [Axiom: AI agents, MCP server and skills](https://axiom.co/docs/console/intelligence/ai-agents-overview), checked September 2026
7. [Axiom: apps and integrations](https://axiom.co/docs/apps/introduction), checked September 2026
8. [Axiom Trust Center](https://trust.axiom.co/), checked September 2026
9. [Axiom for Vercel](https://axiom.co/vercel), checked September 2026

## Other comparisons

- [Maple vs Datadog](https://maple.dev/compare/datadog.md)
- [Maple vs Grafana](https://maple.dev/compare/grafana.md)
- [Maple vs New Relic](https://maple.dev/compare/new-relic.md)
- [Maple vs Dash0](https://maple.dev/compare/dash0.md)
- [Maple vs SigNoz](https://maple.dev/compare/signoz.md)