Unikraft
Also known as: Unikraft Cloud
Cloud platform that runs any Dockerfile as a microVM with hardware isolation, millisecond cold starts and scale to zero, used for agent sandboxes, headless browsers and MCP servers.
Unikraft Cloud runs any workload built from a Dockerfile as a minimal Linux microVM with hardware level isolation. It is a commercial cloud platform from Unikraft Inc., not the open source unikernel project at unikraft.org, and not a unikernel product. Instances cold start in under ten milliseconds, scale to zero when idle and wake in milliseconds on the next connection, and more than 100,000 isolated instances can run on one server.
For agents, Unikraft hosts sandboxes for untrusted AI generated code, headless and headful browsers with GPU rendering, remote desktops, and MCP servers that wake when an agent calls them. Snapshots initialize a workload ahead of time, so new instances boot warm from a template. A live VM and its children can be forked in about ten milliseconds, and managed volumes keep data. The customer brings its own browser, tools and agent logic. Unikraft provides the isolated machines and the lifecycle around them.
Unikraft Cloud runs in named metros in the US, Germany and Singapore. It can also be deployed dedicated, on premises or as BYOC on bare metal or AWS, GCP and Azure VMs, or added as a node to an existing Kubernetes cluster. The company holds SOC 2 Type II and is HIPAA compliant with BAAs available, and DEFION Security has run an independent penetration test of the platform.
Plans run from a free Hobby tier to Team at $39 a month and Pro at $199 a month, with custom Enterprise pricing.
Vendor details
Canonical URL
https://unikraft.com
Category
Agent infrastructure
Funding status
Independent. Unikraft Inc. was founded in 2022 by Felipe Huici (CEO), Simon Kuenzer (CTO) and Alexander Jung (CPO).
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Unikraft Cloud runs any Dockerfile workload and is driven through its REST API and the unikraft CLI. It joins Kubernetes as a node in EKS, GKE, AKS or the customer's own cluster, and works with Karpenter. Metrics export in Prometheus format, and the platform deploys on bare metal or on AWS EC2, GCP Compute Engine and Azure VMs.
In practice
Your agent calls a dozen MCP tools, most of them idle most of the time. You run each as a Unikraft microVM that wakes in milliseconds when called and costs nothing between calls.
Your browser agents need a fresh, isolated browser per session without waiting for Chrome to start. You snapshot a warmed browser into a Unikraft template and boot every session from it in milliseconds.
Your security team will not let AI generated code run in shared containers. You run each execution in its own Unikraft microVM with hardware isolation, on premises if the data cannot leave your network.
Sources & related URLs
Related / legacy domains
Agentic Index coverage score
6.0 / 14 capabilities · 43%
| Integrations & Tool Calling | Partial |
|---|---|
|
Agents' tools run on Unikraft instead of coming from it. MCP servers run as microVMs that wake when an agent calls them, alongside webhook receivers and API gateways, and workloads connect to hyperscaler services such as S3. One customer keeps its credentials outside the VM. Unikraft names no connector catalog, OAuth actions or credential brokering of its own, so the read and write integrations are the customer's code. Sourceunikraft.com/docs/use-cases/mcp-serversread 2026-09-22 |
|
| Workflow Orchestration | Not documented |
|
The platform schedules and scales microVMs, with autoscale, load balancing and forking a live VM for parallel sub agents. It does not sequence, branch, retry or route an agent's steps, and that logic lives in the customer's code. Unikraft offers no workflow logic that mixes deterministic nodes with agent steps. Scheduling across metros decides where compute runs, not how a workflow proceeds. Sourceunikraft.com/llms-full.txtread 2026-09-22 |
|
| Knowledge Grounding & RAG | Not documented |
|
The platform runs the customer's workloads, including databases and MCP servers, but keeps no retrieval structure over the customer's content for an agent to query. Volumes and databases hosted on it are the customer's own storage. There are no sources that Unikraft indexes, refreshes or permissions. Sourceunikraft.com/llms-full.txtread 2026-09-22 |
|
| Human Oversight & Guardrails | Not documented |
|
Unikraft names no approval step, consent checkpoint or escalation to a person for the work it runs, and no place to insert one. The network shield and relay, an Enterprise feature, routes an instance's traffic through a relay VM, and delete locks stop instances being removed. The developer configures both for isolation and protection, and neither has a person review an agent's action. Sourceunikraft.com/docs/features/network-shieldread 2026-09-22 |
|
| Security, Identity & Governance | Full |
|
The company holds SOC 2 Type II against the Security criteria and is HIPAA compliant, with BAAs available. An independent penetration test by DEFION Security found no exploitable cross VM, hypervisor, filesystem or host agent issues. Alongside these, every workload is its own Firecracker based microVM with per tenant filesystem boundaries, delete locks protect instances from removal, and on Enterprise a network shield and relay route an instance's traffic through a relay VM. Unikraft names no SSO, roles or identity provider support. Sourceunikraft.com/llms-full.txtread 2026-09-22 |
|
| Observability & Auditability | Partial |
|
Logs and metrics follow the instance, not the agent. Every user gets full stdout and stderr logs, paying customers get Prometheus metrics, and a metrics API reports CPU time, memory, network I/O, connection counts and scale to zero wake latency per instance. The metrics export to monitoring systems. Inspecting prompts, tool calls, retrieved knowledge and outputs step by step is left to the customer's own code. Unikraft names no audit log separate from the runtime logs and no retention by plan. Sourceunikraft.com/docs/platform/metricsread 2026-09-22 |
|
| Memory & State Persistence | Partial |
|
A workload's state carries across sessions. Instances snapshot and resume in milliseconds with their state intact, a live VM and its children fork in about ten milliseconds, managed volumes persist data, and paid plans add snapshot history and rollback. The agent's code reads that restored state to carry on. Volumes are storage the code reads, and restored process state belongs to the machine. The customer can resume, roll back or delete that state but cannot review or edit what it holds. Unikraft offers no separate memory layer for agents. A browser session ROM keeps its own cookies and profile, much like a saved login. Sourceunikraft.com/docs/features/snapshotsread 2026-09-22 |
|
| Deployment & Data Residency | Full |
|
Customers choose the named metros where Unikraft Cloud runs. Dallas, Frankfurt, San Francisco, Singapore and Washington DC are the defaults, and the list is set per account. Deployment can be hosted, dedicated hosted, on premises or BYOC, on bare metal or on AWS EC2, GCP Compute Engine or Azure VMs. A Kubernetes integration adds Unikraft Cloud as a node in EKS, GKE, AKS or the customer's own cluster. Sourceunikraft.com/docs/platform/metrosread 2026-09-22 |
|
| Prebuilt Agents / Templates / Packs | Not documented |
|
Templates here are snapshots of the customer's own warmed up workload that new instances boot from, not ready made starting points from Unikraft. Unikraft covers patterns for sandboxes, headless browsers, MCP servers and remote desktops, but ships no working agents, production ready templates or role specific agents for a customer to adopt. Sourceunikraft.com/docs/features/on-demand-templatesread 2026-09-22 |
|
| Triggers & Channel Coverage | Full |
|
Incoming connections wake work without the customer's code starting it. An instance scaled to zero consumes nothing until the next connection arrives. The platform then boots or resumes it in milliseconds and answers, so an MCP server wakes when an agent calls it and a webhook receiver wakes on the incoming event. A tutorial walks through scale to zero triggers. Unikraft names no scheduler. Sourceunikraft.com/docs/features/scale-to-zeroread 2026-09-22 |
|
| Model Flexibility & Routing | Not documented |
|
Any Dockerfile workload runs here, and nothing in the platform is specific to language models. There is no model selection, routing, provider keys or guidance for connecting the customer's own models. GPUs appear only for rendering in headful browsers. The platform imposes no model of its own either. Sourceunikraft.com/llms-full.txtread 2026-09-22 |
|
| APIs / SDKs / MCP Extensibility | Full |
|
From outside, a platform REST API drives Unikraft Cloud, covering instances, templates, volumes, services and metrics. The unikraft CLI adds a raw API passthrough, and a Kraftfile specification and Kubernetes and Karpenter integrations round it out. Unikraft names no SDK or MCP server of its own. Some newer features, such as plugins and the network shield, are available only through the API for now. Sourceunikraft.com/docs/llms.txtread 2026-09-22 |
|
| Testing, Debugging & Optimization | Not documented |
|
Tests get somewhere to run, but Unikraft does not test agents. Forking and the build and test environments pattern give the customer disposable, parallel VMs. The platform has no fixtures, scoring, quality gates or comparison of agent runs. Sourceunikraft.com/docs/llms.txtread 2026-09-22 |
|
| Browser / Computer-use | Partial |
|
Browser and desktop agents get their machines here. Every browser session can run in its own disposable headless or headful microVM with GPU rendering, booted from a snapshotted template, and whole desktop sessions snapshot after login and resume on connect. Browser agent companies run on it. The platform works with any browser, so the customer brings the browser and drives it with its own tooling. Unikraft offers no API of its own for screenshots, clicks or typing. It supplies the environment for operating web interfaces and desktops, and the customer supplies the control. Sourceunikraft.com/docs/use-cases/headless-browsersread 2026-09-22 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Product updates
Release 14 of Unikraft Cloud adds instance plugins and a first sandbox plugin that gives a running microVM remote command execution, an interactive shell and file transfer through new unikraft instance exec, shell, copy, write and read commands. Plugins load when an instance is created, run or edited, and the same release lets microVMs grow and shrink their memory and vCPUs while they run.
Bears on: Security / enterprise
View sourceUnikraft launched Release 12, codenamed Thebe, which introduces NVIDIA GPU support for running GPU-accelerated inference and headful browser agents. The release also adds cluster-wide instance templates and advanced image control features like image pinning, strict authorization for shared private registries, and pull timeouts.
Bears on: Deployment / data residency
View sourceUnikraft detailed its agent infrastructure platform architecture, optimized for running AI sandboxes at scale. The platform leverages hardware-isolated microVMs to deliver 10-millisecond cold starts and stateful scale-to-zero capabilities, allowing agents to pause and resume exactly where they left off. It also features a dedicated network shield that enforces outbound traffic policies and isolates credentials from the agent environment.
Bears on: Agent capability
View sourcePricing
Hobby free · Team from $39/mo · Pro $199/mo
compute usage (instances, VM runtime, resources)
What is public
Unikraft publishes four tiers (Hobby free, Team $39, Pro $199, Enterprise custom) with instance limits and features per tier.
Billing mechanics
A flat monthly plan that sets how many instances can run and stand by, with scale to zero so idle instances cost nothing; Enterprise is contracted.
Cost watchouts
Plans cap running and standby instances by tier, so fleet size can force an upgrade; Enterprise dedicated, on prem and BYOC are quoted.
Variable cost rationale
Compute infrastructure billed on usage, so total cost scales with the number of instances, VM runtime, and resource consumption rather than a fixed plan fee. Scale to zero reduces idle cost.
Additional watchouts
Check the running and standby instance limits for your fleet size before choosing a tier.
Sales call required
Mixed (some tiers require a call)
Free / trial
Hobby plan free forever, no credit card required
Lowest paid plan
Team $39/month
Key ambiguities
Plan prices are public, but whether compute beyond each plan's limits is billed by usage, and at what rates, is not stated.
Missing data
Usage rates beyond plan limits and the full Team plan limits.
Related vendors
- AgentOps — Agent observability and debugging platform: open source SDKs trace…
- AgentPhone — AgentPhone gives the AI agents a team builds their own US or…
- Agno — Python agent framework and AgentOS runtime (formerly Phidata) for…
- AIsa — Resource and payment gateway for AI agents, with one key to 110+…
- Albato — Albato is a no code integration platform that sells an embedded,…
- AlphaBitCore — AI control plane for regulated financial firms: one gateway enforces…
Alternatives to Unikraft
The closest documented capability profiles to Unikraft among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Modal8.0 / 14Adds documented Workflow Orchestration and Prebuilt Agents, Templates & Packs, among others
- Exa6.5 / 14Adds documented Prebuilt Agents, Templates & Packs and Model Flexibility & Routing
- Rime3.5 / 14A lighter documented profile than Unikraft
- Speechmatics3.5 / 14A lighter documented profile than Unikraft
- Daytona8.0 / 14Adds documented Prebuilt Agents, Templates & Packs and Model Flexibility & RoutingUnikraft vs Daytona →
- E2B8.0 / 14Adds documented Prebuilt Agents, Templates & Packs and Model Flexibility & RoutingUnikraft vs E2B →
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded