About PandaProbe Cloud
Overview
PandaProbe Cloud is a managed platform for tracing, evaluating, and monitoring AI agents, built by Chirpz AI as the hosted counterpart to its open-source stack. It runs trace ingestion, storage, dashboards, and the evaluation models on PandaProbe's own infrastructure, so engineering teams operate agent fleets without maintaining observability tooling of their own. One product covers hobbyist builders, small development teams, and large organizations with enterprise security requirements.
Setting up Cloud takes minutes rather than the days to weeks that self-hosted deployments demand. The managed edition includes SSO, role-based permissions, a built-in eval scheduler, managed storage and retention, auto-scaling, and SLA-backed support. Deployment options range from a free Hobby plan to a custom Enterprise agreement with hybrid or self-hosted hosting.
Key Benefits
- Hosted trace ingestion and dashboards remove server provisioning, maintenance, and capacity planning from your team.
- Managed LLM-as-judge and embedding models run evaluations without external API keys.
- Auto-scaling covers traffic spikes, growing teams, and enterprise-scale trace volumes.
- Role-based access control and SSO meet enterprise security requirements as teams grow.
- A built-in eval scheduler runs daily, hourly, or custom cron evaluations against production traffic.
- Dedicated support channels and SLA guarantees back paid plans so teams ship with confidence.
How It Works
Teams sign up for PandaProbe Cloud and install the PandaProbe skill with npx skills add chirpz-ai/pandaprobe-skills using a terminal or a coding agent, which wires the project to the hosted platform. Traces then flow into managed dashboards, the scheduler runs judge and embedding evaluations on the chosen cadence, and reviewers add human annotation to label results.
Use Cases
- AI engineers trace production agent runs in hosted dashboards to find where failures originate.
- Evaluation engineers schedule hourly or daily judge runs against live traffic to catch regressions early.
- Small development teams start on the Pro plan and pay as they go when trace volume grows.
- Scaling startups use high rate limits, a private Slack channel, and retention management on the Startup plan.
- Enterprise platform teams add custom SSO, hybrid or self-hosted hosting, team trainings, and SLA-backed support.
- Hobbyists begin free with 100 base trace ingestions, 100 trace eval runs, and human annotation each month.
Why Choose This Product
Cloud removes the operational work of self-hosted tracing: no servers to provision, no eval API keys to supply, and no capacity planning. Setup drops from days or weeks to minutes, PandaProbe runs the evaluation models and storage, and paid tiers add dedicated support with an SLA. Teams that prefer to own their infrastructure can deploy the Apache 2.0 open-source edition or request an Enterprise hybrid deployment.
PandaProbe Cloud Pros & Cons
- Cloud setup takes minutes versus days to weeks for self-hosted
- Hobby plan is free forever with no credit card required
- Managed eval LLM and embeddings need no external API keys
- Built-in eval scheduler runs hourly, daily, or custom cron
- Enterprise adds hybrid hosting, custom SSO, and a support SLA
- Hobby plan caps at 1 seat and 100 base trace ingestions per month
Key Features
Hosted Trace Ingestion
Cloud ingests, stores, and displays agent traces in managed dashboards with no servers to provision or scale.
Managed Eval Models
PandaProbe Cloud runs the LLM-as-judge and embedding models for evaluations, so no external API keys are required.
Eval Scheduler
A built-in scheduler runs evaluations daily, hourly, or on custom cron schedules against production traffic.
Auto-Scaling
The platform absorbs traffic spikes, growing teams, and enterprise volumes without manual capacity planning.
SSO & Permissions
Role-based access control and SSO cover growing engineering teams and enterprise security requirements.
Human Annotation
All plansTeams add human annotation to review and label agent runs alongside automated evaluations.
Data Retention Controls
Startup+Data retention management is available on the Startup and Enterprise plans for control over stored traces.
Skill Installer
A skill installer connects a project to Cloud from a terminal or coding agent with npx skills add chirpz-ai/pandaprobe-skills.
Pay-As-You-Go Overages
Pro and Startup include monthly base allowances and bill usage beyond those allowances pay-as-you-go.
SLA & Support
Paid plans add email, private Slack, or dedicated support channels, with SLA guarantees and a dedicated engineering team on Enterprise.
PandaProbe Cloud Pricing
View full pricing โ- 100 base trace ingestion per month
- 100 trace eval runs per month
- 10 session eval runs per month
- Human annotation
- 1 seat
- Community support via GitHub
- 5,000 base traces per month, then pay-as-you-go
- 5,000 trace eval runs per month, then pay-as-you-go
- 100 session eval runs per month
- 2 seats
- Email support
- 50,000 base traces per month, then pay-as-you-go
- 50,000 trace eval runs per month
- 1,000 session eval runs per month
- 10 seats
- High rate limits
- Private Slack channel and data retention management
- Alternative hosting options (hybrid and self-hosted)
- Custom SSO
- Access to dedicated engineering team
- Support SLA
- Team trainings and architectural guidance
- Unlimited seats
| Feature | Hobby Free | Pro $29/mo | Startup $299/mo | Enterprise Custom |
|---|---|---|---|---|
| 100 base trace ingestion per month | ||||
| 100 trace eval runs per month | ||||
| 10 session eval runs per month | ||||
| Human annotation | ||||
| 1 seat | ||||
| Community support via GitHub | ||||
| 5,000 base traces per month, then pay-as-you-go | ||||
| 5,000 trace eval runs per month, then pay-as-you-go | ||||
| 100 session eval runs per month | ||||
| 2 seats | ||||
| Email support | ||||
| 50,000 base traces per month, then pay-as-you-go | ||||
| 50,000 trace eval runs per month | ||||
| 1,000 session eval runs per month | ||||
| 10 seats | ||||
| High rate limits | ||||
| Private Slack channel and data retention management | ||||
| Alternative hosting options (hybrid and self-hosted) | ||||
| Custom SSO | ||||
| Access to dedicated engineering team | ||||
| Support SLA | ||||
| Team trainings and architectural guidance | ||||
| Unlimited seats |
Pricing extracted from the product website and may change. Check the source for current details.
Frequently asked questions about PandaProbe Cloud
What is PandaProbe Cloud?
PandaProbe Cloud is a fully managed platform from Chirpz AI that provides full-stack tracing, evaluation, and monitoring for AI agents. It runs trace ingestion, storage, dashboards, and the evaluation models for you, so your team ships better agents without the operations overhead.
Is there a free plan?
Yes. The Hobby plan costs $0 forever and includes 100 base trace ingestions per month, 100 trace eval runs, 10 session eval runs, human annotation, one seat, and community support via GitHub. No credit card is required.
Do I need my own API keys to run evaluations?
No. PandaProbe Cloud runs the evaluation LLM-as-judge and embedding models for you, so no external API keys are required. Self-hosted deployments instead bring their own eval LLM and embeddings.
How is Cloud different from self-hosting?
Cloud is hosted and managed: setup takes minutes instead of days to weeks, and hosted trace ingestion, managed eval models, a built-in scheduler, storage, SSO, and SLA support come included. Self-hosting runs the Apache 2.0 open-source stack on your own infrastructure with community support.
How does Cloud pricing work?
Each plan includes monthly base allowances for base trace ingestion, trace eval runs, and session eval runs. Pro and Startup bill usage beyond those allowances pay-as-you-go, and Enterprise pricing is custom.
How quickly can I get started?
Setup takes minutes. Sign up for PandaProbe Cloud, then use the installer with a terminal or coding agent โ npx skills add chirpz-ai/pandaprobe-skills โ to connect your project.

