Kortix
Free tierThe AI Command Center for Your Company
Free tier available·All audiences·Powered by Model-agnostic: supports Anthropic Claude, OpenAI, Google Gemini, DeepSeek, GLM, and any OpenAI-compatible endpoint; bring your own API key or ChatGPT subscription·API available·Open source
Key strengths
Open-source AI Operating System with full ownership via a single git repoModel-agnostic: supports any major LLM provider with bring-your-own-keySelf-hostable (VPC, on-prem, or managed cloud)3,000+ app connectors plus MCP, OpenAPI, Postman, GraphQL, and raw HTTPIsolated per-session Linux sandboxes for safe parallel agent executionMulti-channel access: Web, Slack, Microsoft Teams, CLI, Mobile, and API/SDK
Free tier + paid plans · from $40 USD/mo
Founded 2024
Self-hostable
Facts verified against kortix.com on
Recent changes
- Plans updated: Free tier: $0/month with 500 credits/month, 1 project, no credit card required. Credits tier: $200/month for 20,000 credits (~1,830 agent-computer hours), unlimited members, unused credits roll over one month. Enterprise: custom pricing with SAML SSO, RBAC, VPC/on-prem, SLA. 1 credit = $0.01; 11 credits = 1 hour of agent compute (2 vCPU / 4 GB / 20 GB).
- Models updated: Model-agnostic: supports Anthropic Claude, OpenAI, Google Gemini, DeepSeek, GLM, and any OpenAI-compatible endpoint; bring your own API key or ChatGPT subscription
Get alerts when Kortix changes
Free. We'll email you when pricing, the free tier, API access or open-source status changes.
No ratings yet
Kortix — Technical Overview
Kortix is an open-source AI Operating System (20,321 GitHub stars) that lets engineering teams deploy, orchestrate, and manage AI agents across an entire company workflow from a single self-owned git repository.
Architecture Highlights
- Model-agnostic runtime: Natively supports Anthropic Claude, OpenAI, Google Gemini, DeepSeek, GLM, and any OpenAI-compatible endpoint. Bring your own API key or ChatGPT subscription — no vendor lock-in.
- Isolated execution environments: Each agent session runs in a dedicated, per-session Linux sandbox (2 vCPU / 4 GB RAM / 20 GB storage), enabling safe parallel agent execution without cross-session interference.
- 3,000+ integrations: Connect via MCP, OpenAPI specs, Postman collections, GraphQL, raw HTTP, or pre-built connectors. First-party integrations include Slack, Microsoft Teams, GitHub, Apollo, and Email.
- Multi-channel access: Agents are reachable through Web UI, Slack, Microsoft Teams, CLI, Mobile, and a full API/SDK.
- Self-hostable: Deploy on VPC, on-premises, or use the managed cloud. Enterprise tier adds SAML SSO, RBAC, and SLA guarantees.
- Open-source & self-owned: The entire platform lives in a single git repo — fork it, audit it, extend it.
Pricing & Compute Model
| Tier | Price | Credits | Compute |
|---|---|---|---|
| Free | $0/month | 500 credits/month | ~45 agent-hours |
| Credits | $200/month | 20,000 credits (~1,830 agent-hours) | Unused credits roll over 1 month |
| Enterprise | Custom | Custom | SAML SSO, RBAC, VPC/on-prem, SLA |
1 credit = $0.01 · 11 credits = 1 hour of agent compute (2 vCPU / 4 GB / 20 GB).
Alternatives
See all alternatives