> ## Documentation Index
> Fetch the complete documentation index at: https://docs.condense.io/llms.txt
> Use this file to discover all available pages before exploring further.
# AI Agents
## **Meet Vapr™ : A Specialized AI Agent for Multi-Layer Data Operations for Condense Platform**
Vapr turns complex technical intent into finished results by orchestrating specialized agent workforce across your data lifecycle, resolving systems instantly through precise execution and deep technical understanding
## **Vapr's Capabilities**
Vapr is an AI agent that functions as an autonomous supervisor and intelligent routing interface, designed to assist users by orchestrating a hierarchical workforce of specialized agents within the Condense platform, where simple tasks are routed for immediate action via tool-specific L1 specialist agents, while complex requirements are delegated to L2 agents for multi-step planning and logical design.
## **What You Can Do with Vapr**
You can use Vapr as your central supervisor to manage your entire operation. By simply stating your goal, Vapr identifies the need and triggers the right agent workforce to handle the heavy lifting.
> No more juggling tools or tracking tasks manually. Vapr routes work, coordinates agents, and ensures completion end-to-end
You can instantly audit Kafka topics, navigate complex schema registries, and retrieve live messages to validate your data flow in real time.
> No more CLI digging or fragmented visibility. Vapr gives you a unified, real-time view of your entire data stream
You can summarize Kubernetes cluster states, inspect service configurations, and aggregate logs across namespaces to keep your environment stable and scalable.
> No more manual audits or scattered logs. Vapr centralizes visibility to monitor and maintain your infrastructure
You can identify the exact point of failure before a crisis starts by correlating Prometheus metrics with system telemetry to deliver a Root Cause Analysis (RCA).
> No more reactive debugging or guesswork. Vapr detects anomalies early and pinpoints root causes instantly
You can plan, architect, and generate production-grade code through the Coding Agent.
> No more tool switching or manual handoffs. Vapr automates workflows from code generation to pull request creation
AI Agents are fully integrated into Condense and help engineering teams reduce operational effort, accelerate troubleshooting, and improve developer productivity.
***
# Available AI Agents
| AI Agent | Purpose |
| :--------------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Kafka Agent** | Retrieves Kafka topics, schemas, consumer groups, metadata, partition information, and recent messages from Kafka. |
| **Kubernetes Agent** | Retrieves Kubernetes resources including services, jobs, Kafka clusters, users, Schema Registry information, and connector logs. |
| **Grafana Monitoring Agent** | Diagnoses Grafana alerts and retrieves monitoring data from Prometheus and Kubernetes to assist with root cause analysis. |
| **Coding Agent** | Generates Condense Applications, Connectors, and Transforms, manages project files, creates automated tests, and integrates with GitHub and GitLab repositories. |
Additional AI Agents will be introduced in future Condense releases.
***
## **How Do These Agents Talk to Vapr™**
Each AI Agent specializes in a particular domain and executes predefined tools to retrieve or process information.
Instead of manually navigating dashboards or infrastructure resources, users can request information from Vapr, which invokes the appropriate tool and returns contextual results.
##### **You provide the goal to Vapr, which functions as an autonomous supervisor to orchestrate the right mix of L1 specialists and L2 architects for the task**
##### **Agents talk to each other through Vapr to share context, such as using a Grafana alert to trigger a Kubernetes scaling action or a Git update**
##### **The Vapr ensures that no agent acts alone; every technical resolution is cross-referenced against system’s state and requires final review of user**
Depending on the agent, supported operations include:
Kafka Agent
Real-time data retrieval from Kafka topics, schema management, and stream configuration
* Topic Insight
* Message Retrieval
* Schema Registry
* Consumer Analytic
Kubernetes Agent
Environment transparency, log aggregation & cluster state reading by retrieving K8s pod, job & service details
* Resource Inspection
* Deep Logging
* Cluster Discovery
* Registry Access
Monitoring Agent
Proactive diagnostics and RCA by analyzing Grafana alerts and retrieving insights from Prometheus & K8s logs
* Automated RCA
* Telemetry Access
* Metric Discovery
* Target Health
Developer Agent
Turning high-level logic into production-ready code and repository management
* Logic & Planning
* Filesystem & Development
* Git Integration
* Direct Push
***
# Benefits
Condense AI Agents help organizations:
* Reduce operational overhead.
* Accelerate troubleshooting.
* Improve platform observability.
* Simplify Kafka and Kubernetes operations.
* Assist developers in building Condense applications.
* Automate repetitive engineering tasks.
* Improve productivity across development and operations teams.
***