Dedicated Firecracker microVMs
Each agent receives its own KVM-isolated virtual machine with allocated CPU, RAM, disk, and network resources. The service describes this as hardware-level isolation rather than a shared-kernel container boundary.
jurniti
Freemium
jurniti é uma plataforma de hospedagem gerenciada que mantém agentes de IA online em microVMs dedicadas do Firecracker. Oferece credenciais de modelo BYOK, painel web, CLI, acesso MCP e armazenamento persistente, além de planos com tarifa fixa ou por uso para execuções contínuas ou pontuais.
O que é jurniti?
jurniti is a managed hosting platform for AI agents. Each agent runs in its own Firecracker microVM with dedicated CPU, memory, disk, and network resources rather than in a shared container slot. Users bring their own agent runtime and model-provider credentials; jurniti supplies the virtual machine, network, provisioning, and billing. The service is intended for agents that should remain available without leaving a laptop running.
Agents can be managed through a browser dashboard or the jurniti CLI, and an agent can access the same platform operations through MCP. The platform supports flat monthly plans for always-on workloads and prepaid On-Demand or Spot usage for burstier workloads. State written under /persist can survive restarts, resizing, and snapshots, and configured agents can be shared or reproduced through templates.
O que jurniti pode fazer?
Bring your own model keys
Users enter their own credentials for supported model providers, including OpenRouter, OpenAI, Anthropic, and Nous Portal. Keys remain inside the agent VM, and jurniti does not add token markups or proxy model requests.
Dashboard, CLI, and MCP control
Provision and inspect agents from the web dashboard, use the CLI for terminal work, CI, and scripts, or mount jurniti through MCP so another agent can use the platform's operations as tools. A REST API with an OpenAPI specification is also documented.
Persistent agent state
Data stored under `/persist`, such as skills, memory, conversation history, plugin state, embeddings, vector stores, scripts, and uploaded files, is retained through restart, resize, and snapshot operations.
Templates and forking
A configured VM can be turned into a public or private template. Forking copies the persist volume into a new paid VM; the harness reinstalls on first boot and then reads the inherited state.
Flexible billing for runtime patterns
Flat monthly plans are available for always-on agents, while prepaid On-Demand and Spot tiers bill for running time. Stopped metered agents do not consume runtime, and Spot capacity can be interrupted or queued when capacity is unavailable.
Casos de uso
“Always-on personal agents”
Host an agent continuously so it remains available through its web interface, assigned public TLS subdomain, webhook, or chat bridge without keeping a personal computer awake.
“Repeatable agent deployments”
Configure one working agent, capture its persistent state in a template, and fork it into separate VMs for new projects or independently configured instances.
“Burst or part-time workloads”
Use On-Demand or Spot credits for agents that run only during scheduled tasks, experiments, or short work sessions instead of paying for a continuously running VM.
“Terminal and automation workflows”
Provision and inspect VMs from scripts, CI, or a local terminal using the CLI. MCP and the REST API provide additional ways to expose platform operations to software and agents.
“Shared agent ecosystems”
Browse public templates or publish a configured setup with a README, tags, and a stable URL so others can reproduce the setup in their own subscription. Forked templates should be reviewed before use because they may contain unfamiliar tools or stale credentials.
Perguntas frequentes
How do I start an agent on jurniti?
You can use the dashboard with magic-link login, add a card, select Add agent, and open the browser terminal. Alternatively, install the CLI on macOS, Linux, or Windows, log in, add a payment method, provision a VM with a harness and plan, and connect. The site describes initial provisioning and connection as taking about five minutes, with agents commonly becoming live in roughly three minutes after provisioning or a template fork.
Which model keys does jurniti use?
jurniti uses a BYOK model. The pricing documentation names OpenRouter, OpenAI, Anthropic, and Nous Portal as supported providers. Users place their keys inside the VM, and jurniti says it does not proxy requests, inspect token counts, or add token markups.
Does jurniti offer a free plan or trial?
No. jurniti requires a card on file and provisions a real, billed microVM. The first subscription instead includes a 30-day money-back guarantee; the service says the first plan can be canceled within that period for a full refund.
What happens when I fork a template?
Forking creates a new VM and subscription. The template's persistent drive is restored, the harness installs fresh on first boot, and the agent then reads the inherited configuration and data. VM memory state is not copied. Review a template before entering your own credentials because its persistent files may contain secrets or unfamiliar code.
Can I use jurniti without installing the CLI?
Yes. Dashboard-only users can provision and operate agents in the browser. The CLI is intended for terminal use, CI, scripts, and MCP, and prebuilt binaries are available for macOS, Linux, and Windows.
Informações rápidas
- Category
- Managed AI agent hosting
- Isolation
- One Firecracker microVM per agent
- Control surfaces
- Web dashboard, CLI, MCP, and REST API
- Supported CLI platforms
- macOS, Linux, and Windows
- Credential model
- Bring your own model-provider keys
- Billing options
- Flat monthly plans, On-Demand credits, and Spot credits
Análise de tráfego de jurniti
Os dados de tráfego são apenas para referência.
- Classificação do domínio
- 3
Alternativas ao jurniti
Experiential Labs
www.experientiallabs.ai
Experiential Labs é um gateway de IA de código aberto que oferece às equipes um único endpoint e uma chave de API para modelos hospedados, chaves dos próprios clientes e modelos locais ou personalizados. Também inclui roteamento, orçamentos, controle de acesso, atribuição e recursos de inteligência opcionais para o tráfego de agentes.
Weave
workweave.ai
Weave é uma plataforma de inteligência de engenharia que mede a produção das equipes, o trabalho assistido por AI e o impacto das revisões de código para equipes que desenvolvem software. Oferece os planos Starter, Pro e Enterprise, além de implantação on-premise para organizações maiores.
port22
www.tryport22.com
port22 é um app para iOS e um agente para macOS que monitoram e controlam sessões de agentes de programação em execução no terminal. Transmite transcrições para o iPhone, avisa quando é necessária uma entrada e oferece suporte a Claude Code, Codex e OpenCode.
AIRUNCODE
airuncode.com
AIRUNCODE é um runtime de agentes local-first para executar agentes de código controlados por voz no seu próprio computador. É voltado a quem busca execução local e acesso a provedores de IA pelos preços originais.
AlphaProxy
alphaproxy.ai
AlphaProxy é um gateway compatível com OpenAI que encaminha solicitações entre mais de 450 modelos e 100 provedores, com um runtime de agentes para trabalhos em várias etapas e fluxos agendados. Um único endpoint oferece acesso a modelos, ferramentas conectadas e resultados verificáveis.
QuotaMint
www.quotamint.com
Uma API para gerenciar planos de clientes, acesso a recursos, créditos e acompanhamento do uso, mantendo a cobrança com um provedor existente ou sistema personalizado. Ajuda produtos SaaS e APIs baseados em consumo a decidir em tempo de execução se permitem o acesso e a registrar o consumo.