Free web
Free source (opens in a new tab)
Month
For: Engineers instrumenting agents
Limits
- 2 team seats
- 10K executions/month
1 more
- 14-day retention
Includes
- Observe
- 1 connected client
1 more
- 1 MCP upstream
CodingDocumented overviewLast checked Sep 29, 2026
Governance and observability platform for AI agents: tracing, PII redaction, budget limits and MCP tool monitoring.
(open full size in a new tab)
(open full size in a new tab)
(open full size in a new tab)
(open full size in a new tab)
(open full size in a new tab)
(open full size in a new tab)Our take: Waxell fits engineering and security teams that need tracing, PII redaction and spend limits over production AI agents, starting on a free tier.
Main limitation: Free tier is limited to 2 seats and 14-day data retention; managed Runtime costs an extra $99 per month on Team.
Waxell fits engineering and security teams that need tracing, PII redaction and spend limits over production AI agents, starting on a free tier. Based on the vendor's documentation, not on a hands-on test.
Also a good fit for
Or if
Prices checked Sep 29, 2026 on the official US pricing page. Official pricing ↗ Prices and limits change often.

Free forever tier at $0.
No trial; free tier available.
Free source (opens in a new tab)
Month
For: Engineers instrumenting agents
Limits
Includes
$49/mo source (opens in a new tab)
Month
For: Production teams
Limits
Includes
$499/mo source (opens in a new tab)
Month
For: Governed-agent platform users
Limits
Includes
Custom pricing source (opens in a new tab)
Custom
For: Regulated enterprises
Limits
Includes
LLM token costs go to your own provider; Runtime add-on +$99/mo on Team. A banner mentions 'Full platform $2,499 to start' without detail.
Not stated.
Waxell is a governance and observability layer for teams running AI agents. It instruments agent libraries, records model calls and workflows, and applies policies such as PII redaction and per-agent budget limits. It also offers an MCP gateway that fingerprints tool descriptions and flags changes, plus endpoint discovery of AI apps. Customers bring their own LLM keys.

You provide
You get back
Scenarios are worked out by our editors from the documentation; the actual result depends on your input and plan.

Auto-instruments 200+ libraries with a two-line setup. Reduces manual tracing code.
Collects OpenTelemetry-native traces of agent and model activity. Fits existing observability stacks.
Detects and redacts PII in real time. Limits sensitive data exposure.
Sets budget limits per agent, user and session. Prevents runaway spend.

Waxell has no public API in its official documentation. SDK-based; no separate API stated. API documentation ↗
Summarized from the vendor's current policies; this is not legal advice and a vendor's badge is not an independent audit.
Waxell says it never touches your tokens; provider keys connect directly.
14, 30, 90 or 365+ days by plan.
US or EU data residency on Enterprise.
Policies: waxell.ai/privacy
"Better suited" is based on documented features and pricing, not on a head-to-head test.
| Alternative | When to choose it | Pricing | Limitation | |
|---|---|---|---|---|
| GitHub Copilot | — | — | Compare |
No user reviews yet.
Used Waxell? Be the first to share your experience.
Free has 2 seats, 10K executions per month, 1 connected client, 1 MCP upstream and 14-day retention. source (opens in a new tab)
On the Team plan, Runtime is an add-on at +$99 per month; it is included in Business with 100 CPU-hours. source (opens in a new tab)
No. The site says you bring your own keys and token costs go to your provider with no markup. source (opens in a new tab)
14 days on Free, 30 on Team, 90 on Business and 365+ on Enterprise. source (opens in a new tab)
Not verified: Plan-level governance policy availability; Annual prices.
Every fact about Waxell above comes from the 3 official pages listed here, last compared on Sep 29, 2026; we have not run Waxell ourselves, so it has no editorial score. How we check and re-check pages.
Article updated Oct 10, 2026; prices checked Sep 29, 2026. Outbound links to vendors are plain links; we do not earn commissions. How we review · Report an error
compare>