NetflixFounded by ex-Netflix engineers
CRVBacked by World Class Investors

Build Ops Agents
that stay on top of problems

Catch bad releases, cloud waste, and reliability issues — from your existing data

Windows coming soon • Free to get started

CONNECT TO DATA WAREHOUSES, OBSERVABILITY PLATFORMS, AND MORE

SlackSlack
DataDogDataDog
New RelicNew Relic
GrafanaGrafana
SnowflakeSnowflake
PostgreSQLPostgreSQL
Google BigQueryBigQuery
Google Cloud MonitoringGCP Monitoring
CloudWatchCloudWatch
CloudTrailCloudTrail
AWS LambdaAWS Lambda
PrometheusPrometheus
ThanosThanos
Cardinal LakerunnerCardinal LakerunnerCardinal Release WatcherCardinal
Release Watcher

WORKFLOWS

Choose from Dozens of Ready-Made
Agentic Workflows!

Production-ready workflows that deploy in minutes and instantly go to work for you!

💸

Underutilized but Expensive Cloud Infrastructure

Spot waste automatically. Reclaim cloud spend with confidence.

GCPGCP
BigQueryBigQuery
KubernetesKubernetes

Your cloud bill keeps climbing, but nobody has time to hunt for idle VMs, forgotten GPU nodes, or oversized databases. This workflow continuously analyzes GCP spend against real utilization, highlighting abandoned resources, stale autoscaling groups, and low-traffic services that cost real money. Operators get a clear, prioritized "fix list" with the context to safely shut things down.

1.2k
🚀

Ship Confidently, by Watching Your Releases like You Would

Every deploy tracked. Every regression caught before users feel it.

GitHubGitHub
GCPGCP
DatadogDatadog
0.2k
🔐

Suspicious Access Key Activity — Automatically Investigated and Contained

Access key created? Your agent has already run the investigation.

CloudWatchCloudWatch
CloudTrailCloudTrail
40
🎧

Correlate Support Ticket Volume Spike with a Bad Release

When tickets surge, your agent finds the release that broke things.

ZendeskZendesk
GitHubGitHub
GrafanaGrafana
SnowflakeSnowflake
9

Build Your Own

Create custom workflows tailored to your unique needs.

FEATURES

Everything you need to build
production-ready agents

From data connection to cloud deployment, Cardinal handles the complexity so you can focus on building.

How It Works: Connect Datasources, Talk to it, Build Reports/Knowledge/Workflows, Deploy to Cloud🔌ConnectDatasources💬Talk to it🔧BuildReports, Knowledge, Workflows☁️Deployto Your CloudProductionReadyDrift DetectionPrompt AutotuningToken Monitoring

Multi-Source Connectivity

Connect to any data source—GitHub, Grafana, Datadog, and more. Your agents work with the tools you already use.

Git-Native Workflows

Version your reports, resources, and workflows. PR reviews, diffs, and rollbacks—all supported out of the box.

Local-First Security

Your data and credentials stay on your machine. Institutional knowledge lives in your own private Git repo.

Composable Architecture

Reuse knowledge assets across agents and teams. Build once, deploy everywhere.

Natural Language Editing

Build and maintain workflows using plain English. Extract workflows from chat threads with one click.

Deploy Anywhere

Run scheduled and event-driven workflows on Google Cloud, AWS, or Azure. Scale from prototype to production.

HOW IT WORKS

From idea to production in minutes

1. Chat & Train Your Agent

What's causing this spike in checkout errors, but let me teach you a couple things about checkout errors and how to read them!
Adding which checkout errors are more critical than others to knowledge base...
Checkout errors over timelast 24h
Adding this to my Saved Reports...
⚙️Extract workflow from this thread

2. Extract workflow

⚙️

Extracted workflow

Right-click chat thread to turn this chat into a reusable incident triage playbook.

1Query logs and metrics across datasources
2Generate a one-click incident report
3Post summary + links into Slack / tickets

3. Deploy to your cloud

AWS logo
AWSDeploy workflow as a native job
Ready
GCP logo
GCPDeploy workflow as a native job
Ready
Azure logo
AzureDeploy workflow as a native job
Ready

Deployed workflow

Running in your cloud with guardrails.

⚙️
⚙️Automated Evals
⚙️Drift Detection
⚙️Workflow Linter
⚙️Token Monitoring
Add Knowledge Snippets
Generate Reports
Extract Workflows
Git Native

🔐 Security that you don't have to
think about

VPN-Native by Design

Cardinal Agent Builder is a VPN-native, edge-local app, that keeps every interaction inside your trusted environment—no data ever leaves your VPN.

VPN-Native

Zero-Trust By Default

Agents use the same authentication your teams already trust—Google OAuth, AWS IAM roles, and GCP service accounts. They inherit the agent developer's permissions, nothing more. No shared secrets, no broad access—just true least-privilege AI enforced at the source.

Zero-Trust Authentication

Bring Your Own Models (BYOM)

Use in-VPC models such as AWS Bedrock, Google Vertex AI, Microsoft CloudFoundry or fallback to Cardinal Provided Models by default.

Your Data, Your LLM-Providers, Your Rules.

AWS BedrockVertex AICloud FoundryOpen Source Models
Aarjav Trivedi

"Chat has become the starting point for most of our day-to-day work, and Cardinal nails the UX. Our teams can explore, ask questions, build knowledge, and when ready, turn that conversation directly into a workflow. All the agent steps, guardrails, and context discovered in chat are carried forward automatically. It makes RAG invisible and lets both technical and non-technical users automate real work."

Aarjav Trivedi

CEO, Ridecell

Syed Masthan

"Engineering leaders spend too much time chasing cloud waste, finding underutilized resources, mapping them to teams, and justifying every dollar. We've started exploring Cardinal for this, and the results are promising. I'm looking forward to deleting the very unpopular but recurring 'Cloud Cost Review' meeting from my calendar."

Syed Masthan

Head of Engineering, Moxxi

Kintan Brahmbhatt

"There is still a lot of toil involved in connecting data from disparate data sources, and building end-to-end workflows with push-button reliability. I am very excited about what Cardinal has built to solve this problem!"

Kintan Brahmbhatt

CEO, Olto

Srinivas Nedunuri

"We've been using Cardinal's Lakerunner to power our Telemetry Data Lake for over a year, and it's become a core part of how we manage customer network logs. Our next step is the one we're most excited about: an intelligent triage workflow that links real-time log signals with historical Zendesk tickets. The promise is huge: context-rich, targeted guidance for our support engineers right when they need it."

Srinivas Nedunuri

Lead DevOps Engineer, Aviatrix

Aarjav Trivedi

"Chat has become the starting point for most of our day-to-day work, and Cardinal nails the UX. Our teams can explore, ask questions, build knowledge, and when ready, turn that conversation directly into a workflow. All the agent steps, guardrails, and context discovered in chat are carried forward automatically. It makes RAG invisible and lets both technical and non-technical users automate real work."

Aarjav Trivedi

CEO, Ridecell

Syed Masthan

"Engineering leaders spend too much time chasing cloud waste, finding underutilized resources, mapping them to teams, and justifying every dollar. We've started exploring Cardinal for this, and the results are promising. I'm looking forward to deleting the very unpopular but recurring 'Cloud Cost Review' meeting from my calendar."

Syed Masthan

Head of Engineering, Moxxi

Kintan Brahmbhatt

"There is still a lot of toil involved in connecting data from disparate data sources, and building end-to-end workflows with push-button reliability. I am very excited about what Cardinal has built to solve this problem!"

Kintan Brahmbhatt

CEO, Olto

Srinivas Nedunuri

"We've been using Cardinal's Lakerunner to power our Telemetry Data Lake for over a year, and it's become a core part of how we manage customer network logs. Our next step is the one we're most excited about: an intelligent triage workflow that links real-time log signals with historical Zendesk tickets. The promise is huge: context-rich, targeted guidance for our support engineers right when they need it."

Srinivas Nedunuri

Lead DevOps Engineer, Aviatrix

Aarjav Trivedi

"Chat has become the starting point for most of our day-to-day work, and Cardinal nails the UX. Our teams can explore, ask questions, build knowledge, and when ready, turn that conversation directly into a workflow. All the agent steps, guardrails, and context discovered in chat are carried forward automatically. It makes RAG invisible and lets both technical and non-technical users automate real work."

Aarjav Trivedi

CEO, Ridecell

Syed Masthan

"Engineering leaders spend too much time chasing cloud waste, finding underutilized resources, mapping them to teams, and justifying every dollar. We've started exploring Cardinal for this, and the results are promising. I'm looking forward to deleting the very unpopular but recurring 'Cloud Cost Review' meeting from my calendar."

Syed Masthan

Head of Engineering, Moxxi

Kintan Brahmbhatt

"There is still a lot of toil involved in connecting data from disparate data sources, and building end-to-end workflows with push-button reliability. I am very excited about what Cardinal has built to solve this problem!"

Kintan Brahmbhatt

CEO, Olto

Srinivas Nedunuri

"We've been using Cardinal's Lakerunner to power our Telemetry Data Lake for over a year, and it's become a core part of how we manage customer network logs. Our next step is the one we're most excited about: an intelligent triage workflow that links real-time log signals with historical Zendesk tickets. The promise is huge: context-rich, targeted guidance for our support engineers right when they need it."

Srinivas Nedunuri

Lead DevOps Engineer, Aviatrix

Join Our Team!

Open Roles

Founding Engineer – Frontend

📍San Francisco Bay Area
🕓Full Time

Founding Engineer – Backend

📍San Francisco Bay Area
🕓Full Time

Founding GTM Leader

📍San Francisco Bay Area
🕓Full Time

Start building agents
in minutes, not months

Download Cardinal for free. Upgrade only when you're ready to collaborate and scale to the cloud.

Free to start
No credit card required