Daytona
Infrastructure that gives AI agents isolated, persistent sandboxes in under 90 milliseconds, with container, VM, Windows and GPU options, desktop computer use and OpenTelemetry tracing.
Daytona is infrastructure for running AI generated code, which it now calls the Agent Cloud. It gives an agent a sandbox, an isolated computer with its own kernel, filesystem, network stack and allocated CPU, memory and disk, created in under 90 milliseconds and driven through SDKs, an API, a CLI or an MCP server. Sandboxes run as Linux containers by default, with Linux VM, Windows and GPU sandboxes also available, and macOS sandboxes on reserved Apple silicon come through its use.computer service with a separate API key.
Sandboxes are persistent by default. Files survive stop and start, running processes survive pause and resume on VM sandboxes, and volumes and mounted storage outlive any one sandbox. Sandboxes can also be snapshotted and forked. A Computer Use API gives agents mouse, keyboard, screenshots and recording on Linux and Windows desktops, with VNC for people. Firewalls on each sandbox restrict outbound traffic, secrets hold credentials, and Git, language servers and SSH are built in. Cursor's Cloud Agents can run their self hosted machines on Daytona sandboxes.
For teams, Daytona adds organizations with roles and granular permissions, SSO through OpenID Connect providers on Enterprise, an audit log, webhooks, and OpenTelemetry tracing of SDK calls and sandbox activity to the customer's own collector. Sandboxes run in shared US and EU regions or on the customer's own runner nodes through Bring Your Own Compute. Daytona says it meets HIPAA, SOC 2 and GDPR standards. Customers it names include Meta, Cohere, Writer, Clay, n8n, Postman, Mintlify and bolt.new.
Pricing is usage based and billed per second, with $200 of free compute to start. Compute costs about five cents per vCPU hour and under two cents per GiB of memory per hour, with GPU sandboxes on demand or preemptible.
From our comparison video
Daytona vs Modal: What Agent Sandboxes Cost on Each
Our 4 minute video puts Daytona side by side with Modal. It covers what each one does and how each is priced, then when to choose which.
Recorded in October 2026. The profile here is live and updates as research is verified.
Vendor details
Canonical URL
https://www.daytona.io
Category
Agent infrastructure
Subcategory
Code execution sandboxes
Funding status
Independent. Daytona Platforms Inc., per the site footer.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
SDKs for Python, TypeScript, Ruby, Go and Java, a REST API with OpenAPI specifications (Daytona, Toolbox and Analytics APIs), a CLI, an MCP server for Claude, Cursor and Windsurf, an agent skill, Git and language server support, webhooks, OpenTelemetry export, and guides for Claude Code, Codex, Devin, Gemini CLI, OpenClaw, OpenCode, LangChain, LangGraph, Mastra, the OpenAI Agents SDK and RL frameworks.
In practice
Your agent writes Python and you need to run it without it touching your servers. You execute each run in an isolated Daytona sandbox, then read the output to validate before showing the user.
Your agent's multi step task keeps rebuilding its environment on every retry. You run it in a Daytona VM sandbox and pause it between steps, so it resumes with its files and running processes intact.
You need to see exactly what your agent did inside its sandbox. You turn on Daytona's OpenTelemetry tracing and send every command and file operation to your own observability stack.
Sources & related URLs
Related / legacy domains
Research sources
Agentic Index coverage score
8.0 / 14 capabilities · 57%
| Integrations & Tool Calling | Partial |
|---|---|
|
An agent in a Daytona sandbox reaches outside systems through what runs inside it. Git operations clone and push repositories, secrets hold credentials and external storage can be mounted. Firewall allow lists decide which hosts the agent may reach. Daytona has no catalog of connectors, OAuth actions or tool gateway of its own. The tools are the customer's code. Sourcedaytona.io/docs/llms.txtread 2026-09-22 |
|
| Workflow Orchestration | Not documented |
|
The platform manages sandboxes. It creates and resizes them, applies lifecycle rules and runs warm pools and fleet scaling. It does not sequence, branch, retry or route an agent's steps. In Daytona's guides, agents built in LangGraph, the OpenAI Agents SDK and other frameworks do that, and Daytona has no workflow logic that mixes fixed steps with agent steps. Sourcedaytona.io/docs/en/sandboxesread 2026-09-22 |
|
| Knowledge Grounding & RAG | Not documented |
|
The agent's code runs on Daytona, but there is no retrieval structure over the customer's content. Volumes and mounted external storage hold files the agent's code reads, and there is no index to query. The language server in a sandbox serves code intelligence for the files there. Nothing in Daytona indexes, refreshes or sets permissions on the customer's sources. Sourcedaytona.io/docs/llms.txtread 2026-09-22 |
|
| Human Oversight & Guardrails | Not documented |
|
Work in a sandbox has no approval step, consent checkpoint or escalation to a person. VNC, a web terminal and SSH let a person watch a sandbox and step in by hand. A firewall on each sandbox (allow lists or block all, with organization tiers that sandbox settings cannot override) restricts what an agent's code can reach. Nothing holds an agent's action for a person to review. Sourcedaytona.io/docs/en/network-limitsread 2026-09-22 |
|
| Security, Identity & Governance | Full |
|
Organizations sign in through OpenID Connect providers such as Google Workspace, Microsoft Entra ID and Okta, with provisioning and membership handled at sign in on the Enterprise plan. Collaborative organizations have owner and member roles plus granular assignments by resource, such as SSO Admin and access to the audit log. API keys carry permissions and scopes, with managed child keys. An audit log covers user and system activity, and each sandbox has firewall allow lists or block all. Daytona says it meets HIPAA, SOC 2 and GDPR standards and has a Trust Center, but it names no report type or auditor. Sourcedaytona.io/docs/en/ssoread 2026-09-22 |
|
| Observability & Auditability | Full |
|
Each step an agent takes in the sandbox is recorded. SDK tracing instruments every Daytona API operation and SDK call from the customer's application, such as each command run, file written or process started. Sandbox telemetry collects traces, logs and metrics from inside the sandbox. Both export over OpenTelemetry to the customer's own collector, and Daytona keeps sandbox telemetry for three days. A separate audit log records user and system activity with the user, action and time, and logs can be streamed. The model's reasoning runs outside Daytona. Audit export is not available yet, and Daytona says it is coming soon. Sourcedaytona.io/docs/en/observability/otel-collectionread 2026-09-22 |
|
| Memory & State Persistence | Partial |
|
Persistence is on by default at three layers. The filesystem survives stop, start and archive. On VM sandboxes, memory (running processes and loaded state) survives pause and resume and hot snapshots. Volumes and mounted external storage outlive any one sandbox. Sandboxes can also be forked, and the agent's code reads the restored state to carry on. Paused process state belongs to the machine, so the customer can resume or delete it but cannot review or edit it. There is no separate memory layer that sets what an agent keeps or for how long. Sourcedaytona.io/docs/en/persistenceread 2026-09-22 |
|
| Deployment & Data Residency | Full |
|
Sandboxes run in two shared regions, the United States and Europe, chosen per sandbox. Under Bring Your Own Compute, an organization instead creates custom regions and attaches its own runner nodes, with an optional proxy and snapshot manager for each region, for control over data locality and compliance. Daytona's control plane manages them, and BYOC is sold on the Enterprise plan. The platform sold today cannot be self hosted from open source. Core development has moved to a private codebase, and the public github.com/daytonaio/daytona repository gets no further updates. Its last release, v0.190.0, stays public under AGPL 3.0. Sourcedaytona.io/docs/en/bring-your-own-computeread 2026-09-22 |
|
| Prebuilt Agents, Templates & Packs | Partial |
|
Ready made starting points for running agents include default sandbox snapshots in several sizes, a Declarative Builder for images and an agent skill for coding agents. More than forty guides run coding agents (Claude Code, Codex, Devin, Gemini CLI, OpenClaw, OpenCode), as well as framework agents and reinforcement learning setups, in sandboxes, and a reference implementation runs Cursor's Cloud Agents on self hosted Daytona sandboxes. Developers build from these environments and walkthroughs. There is no browsable catalog of agents a customer can adopt. Sourcedaytona.io/docs/en/agent-skillsread 2026-10-23 |
|
| Triggers & Channel Coverage | Partial |
|
Events go outward. Webhooks notify the customer's endpoint when sandboxes are created or change state and when snapshots and volumes change. Automated lifecycle rules stop, pause, archive or delete idle sandboxes. An agent's work starts only when the customer's code starts it. There is no schedule, and a stopped sandbox does not wake on an inbound request. Sourcedaytona.io/docs/en/webhooksread 2026-09-22 |
|
| Model Flexibility & Routing | Full |
|
The customer chooses the model. Guides run agents on Claude, OpenAI Codex and the Agents SDK, Gemini, Kimi and AWS Kiro, and on open models through OpenCode and other frameworks. GPU sandboxes serve the customer's own LLMs with vLLM or SGLang. Daytona imposes no model, and customers bring their own providers and keys, with guides on setting them up. Sourcedaytona.io/docs/llms.txtread 2026-09-22 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
Outside callers reach Daytona through SDKs in five languages (Python, TypeScript, Ruby, Go and Java), a REST API with published OpenAPI specifications, and a CLI. The specifications cover the Daytona API, the Toolbox API inside the sandbox and an Analytics API. An MCP server lets Claude, Cursor and Windsurf manage sandboxes, and an agent skill serves coding agents. Sourcedaytona.io/docs/llms.txtread 2026-09-22 |
|
| Testing, Debugging & Optimization | Not documented |
|
Evaluations and reinforcement learning run on Daytona, but it does not score them. Its guides run training with HUD, TRL, veRL and OpenEnv in sandboxes, along with a benchmark agent across several languages. The datasets, reward functions and scoring belong to those frameworks. Daytona offers no fixtures, scoring or quality gates for a customer's agent workflows, and no way to compare agent runs. Sourcedaytona.io/docs/llms.txtread 2026-09-22 |
|
| Browser & Computer Use | Full |
|
The Computer Use API gives an agent programmatic control of a desktop in the sandbox, from mouse and keyboard to screenshots, screen recording and display operations, on Linux and Windows. VNC lets a person watch or take over. The desktop runs remotely on Daytona's infrastructure. macOS sandboxes run on Mac hardware reserved for the customer through use.computer, with their own API key, and the Daytona SDKs, CLI and API do not manage them. Sourcedaytona.io/docs/en/computer-useread 2026-10-23 |
|
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
Daytona released the SecretService API for organization-scoped secrets in agent sandboxes. Secrets are injected as opaque placeholders in environment variables, and the real plaintext is substituted only at the network egress layer, and only for allowed hosts.
Bears on: Security / enterprise
View sourcePricing
Usage based · $200 free compute · vCPU $0.0504/h, memory $0.0162/GiB-h
Per second compute and memory, plus metered storage
Included quota
$200 in free compute (not for GPU sandboxes) and the first 5 GiB of storage free. Beyond that, reserved vCPU, memory and disk bill per second by sandbox state.
What is public
Daytona publishes a full per resource rate card for CPU, memory, storage, Windows and GPU sandboxes, with $200 of free compute to start.
Billing mechanics
Sandboxes bill per second for the resources they reserve, by lifecycle state: vCPU, memory and disk while starting or started, disk only while paused or stopped, nothing once deleted. Credits sit in an organization wallet with free and paid balances.
Cost watchouts
Stopped and paused sandboxes still bill for reserved disk; GPU sandboxes cost far more than CPU and cannot use free credits; thousands of concurrent sandboxes for RL or evals add up quickly.
Variable cost rationale
Cost scales directly with sandbox runtime, memory, storage and concurrency, and these workloads expand fast as teams add more agents, longer sessions and parallel runs. A running sandbox keeps billing compute until it stops, by default after 15 minutes idle, and a stopped sandbox still bills for its disk.
Additional watchouts
Delete or archive sandboxes you no longer need, since stopped and paused sandboxes keep billing for disk. Budget GPU work separately from the free credit.
Overage / add-ons
Pay as you go from the organization wallet: reserved vCPU, memory and disk bill per second while started; paused sandboxes bill disk only; deleted sandboxes stop billing.
Sales call required
No, self serve available
Free / trial
$200 in free compute, no credit card required; first 5 GiB of storage free; free credits not usable on GPU sandboxes
Lowest paid plan
Pay as you go usage; no fixed monthly plan
Commercial notes
Sold self serve with free compute and per second billing; Enterprise adds SSO, audit logs, BYOC custom regions and larger limits. A startup program is offered.
Key ambiguities
Total cost depends on reserved resources, how long sandboxes stay started, paused or stopped, concurrency and GPU use.
Cancellation / refund
Self serve, pay as you go usage with no fixed contract on the standard plan. Enterprise terms, including Bring Your Own Compute, are arranged with sales.
Support SLA / resale
Standard usage is self serve. Enterprise adds SSO, audit logs, Bring Your Own Compute and larger limits, arranged through sales.
Missing data
Enterprise pricing is not public.
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 Daytona
The closest documented capability profiles to Daytona 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.
- Firecrawl8.5 / 14Adds documented Human Oversight & Guardrails
- E2B8.0 / 14Fuller documented coverage on Integrations & Tool Calling and Triggers & Channel CoverageDaytona vs E2B →
- Modal8.0 / 14Adds documented Workflow OrchestrationDaytona vs Modal →Video
- Exa6.5 / 14Fuller documented coverage on Triggers & Channel Coverage
- Bright Data8.0 / 14Adds documented Human Oversight & Guardrails
- Hyperbrowser8.0 / 14Adds documented Workflow Orchestration and Human Oversight & Guardrails
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded