#4949th of 102 in Agent platforms
kagent
Kubernetes-native framework for defining and running AI agents as custom resources
- Stars
- 4.4k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- Go
2,687 on 25 Apr 20264,417 on 14 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
kagent defines agents, agent templates, model configs and remote MCP servers as Kubernetes custom resources. A controller reconciles them and runs agents on Go or Python ADKs, Codex, Claude or custom runtimes, with sessions and task history stored in PostgreSQL. It ships a web UI, a CLI, and an MCP server with tools for Kubernetes, Istio, Helm, Argo, Prometheus, Grafana and Cilium.
Who it is for: Platform and SRE teams running AI agents on Kubernetes
Strengths
- Agents, tools and model configs are YAML resources, fitting kubectl and GitOps workflows
- Bundled MCP tools for Kubernetes, Istio, Helm, Argo, Prometheus, Grafana and Cilium
- Sessions persist in PostgreSQL, with suspend, resume, checkpoint and fork
- OpenTelemetry tracing for agent, model and tool execution
Weaknesses
- Requires a Kubernetes cluster and PostgreSQL; no standalone mode described
- Model provider support depends on the chosen harness
- README says the project is in active development
- No hosted demo linked in the README
What it needs
- Needs Kubernetes, PostgreSQL
- Models: OpenAI, Azure OpenAI, Anthropic, Google Vertex AI, Ollama
What changed
“refactor(session)!: remove public suspend and resume controls by @EItanya in https://github.com/kagent-dev/kagent/pull/3409”
“fix(deps): bump Go to 1.27.3 for security fixes by @EItanya in https://github.com/kagent-dev/kagent/pull/3304”
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Agent platforms
See all 102| Rank | Project | Score |
|---|---|---|
| 1 | AutoGPTRole-based AI agents coordinated by one lead agent, plus a visual workflow builder | 95 out of 100 |
| 2 | daily_stock_analysisLLM-driven daily analysis of watchlist stocks across six markets, pushed to chat apps | 95 out of 100 |
| 3 | LangflowVisual builder for AI agents and workflows, served as API or MCP server | 94 out of 100 |
| #4 | deer-flowAgent harness running sub-agents, memory and sandboxes with pluggable skills | 93 out of 100 |
| #5 | DifyVisual builder for LLM workflows, RAG pipelines and agents, with APIs | 92 out of 100 |