Hosted desktop computer-use sandboxes
Give agents a full GUI desktop (mouse/keyboard/screenshot), not just a headless browser.
At a glance — Oct 2026
Updated 2026-10-27 · forward this page: /jtbds/desktop-computer-use#briefScores as of 2026-09-14 (43 days old; Depot marks scores stale after 30 days). The pick is still Depot's current editorial pick; it has not been re-reviewed since. Prices below carry their own dates.
- Depot Pick
- E2B · 8.56/10 for this job (curated estimate, high confidence)
capability desktop APIs + isolation. E2B Desktop SDK is the programmable Linux default; Cua for multi-OS fleets; Scrapybara for turnkey hosted desktops.
- Runner-up
- Scrapybara · 7.58/10
- When not to use the pick
- Need Windows/macOS/Android fleets or local OSS sandboxes — Cua; need turnkey hosted Windows sessions — Scrapybara; browser-only stealth — Anchor/Hyperbrowser; real iOS/Android silicon — BrowserStack / AWS Device Farm (see mobile-computer-use).
- Lock-in / migration
- The pick has lower stored lock-in risk than runner-up Scrapybara (4 vs 4.8). Self-host fallback listed for this job: OSWorld.
- Sourced price (no comparable unit)
Pick: no sourced price in this job's unit (no_matching_sourced_price).
No cheapest SKU (no_comparable_unit_for_job). No price was invented.
- Free option
- OSWorld (oss) — source, as of 2026-10-25
- Failure modes to watch
- Benchmark protocol mismatch: Store assisted-or-not, subset, and harness version beside every cited score. Refuse a ranking that mixes those fields.
- Collateral side effects: Snapshot the workspace or app state, then diff it against an allowlist. A passing check plus an unexpected diff is a failure.
- Confirmation skip: Gate those tool names on an approval id recorded before the call. A missing approval is a failure even if the downstream tool returns success.
- GUI grounding miss: After each GUI action, compare the resulting screen or accessibility tree to the control the plan named. Stop when they differ.
- Partial tool success: Require a per-item status and a failure count. Success for the task requires the failed count to be zero or explicitly handed to a retry queue.
- Prompt injection: Compare each tool call to the original user task. If the call only makes sense given text that arrived from a fetched page, file, or message, stop and ask. Record which source the instruction came from. This catalog does not store payloads or attack steps.
Cite: Agent Provider Depot, Depot Pick for desktop-computer-use: e2b, scores as of 2026-09-14. Agents: GET /api/decide?jtbd=desktop-computer-use
Free Hosted desktop computer-use sandboxes options · for computer-use agents
E2B
capability desktop APIs + isolation. E2B Desktop SDK is the programmable Linux default; Cua for multi-OS fleets; Scrapybara for turnkey hosted desktops.
Pricing (directional): Firecracker microVM sandboxes billed per second; templates; code interpreter for agents.
Lock-in risk: higher = more lock-in (worse). Overall already inverts this dimension.
Depot default code sandbox for agents. Use for untrusted code; Daytona when you need persistent devworkspaces. Also the depot default Linux desktop sandbox for computer-use (e2b-desktop + VNC stream) when you already standardize on E2B.
Quote for Hosted desktop computer-use sandboxes
Depot Pick stays E2B. Prices below are sourced SKU rows only (as of 2026-10-27). Input and output are not blended. GET /api/quote?jtbd=desktop-computer-use
Nothing was compared. no comparable sourced unit for this job; chat input $/1M and other units were not borrowed pickIsPriceWinner is null.
- 1 GiB memory usd_per_gib_second $0.0000045 source as of 2026-10-03
- 1 vCPU usd_per_vcpu_second $0.000014 source as of 2026-10-03
- 2 GiB memory usd_per_sandbox_second_ram_charge $0.000009 source as of 2026-10-03
- 2 vCPU usd_per_sandbox_second_vcpu_charge $0.000028 source as of 2026-10-03
- 4 GiB memory usd_per_sandbox_second_ram_charge $0.000018 source as of 2026-10-03
- 4 vCPU usd_per_sandbox_second_vcpu_charge $0.000056 source as of 2026-10-03
- 6 vCPU usd_per_sandbox_second_vcpu_charge $0.000084 source as of 2026-10-03
- 8 GiB memory usd_per_sandbox_second_ram_charge $0.000036 source as of 2026-10-03
- 8 vCPU usd_per_sandbox_second_vcpu_charge $0.000112 source as of 2026-10-03
- Enterprise minimum usd_per_month $3000 source as of 2026-10-03
- Pro+ concurrency usd_per_month $500 source as of 2026-10-03
- Pro++ concurrency usd_per_month $1000 source as of 2026-10-03
- Pro usd_per_month $150 source as of 2026-10-03
3 of 5 candidates have no sourced price row. 5 are excluded from the Matching price comparison (none). no comparable sourced unit for this job; chat input $/1M and other units were not borrowed
Runners-up
Scrapybara
Clear hosted desktop provider. E2B Desktop remains depot default if you already run E2B; Scrapybara when you…
Cua
Best Scrapybara/E2B Desktop peer when you need Windows+macOS+Android fleets or local OSS sandboxes. E2B…
Anchor Browser
Strong when bot-detection and authenticated computer-use matter. A/B vs Browserbase/Kernel.
When NOT to use the top pick
Need Windows/macOS/Android fleets or local OSS sandboxes — Cua; need turnkey hosted Windows sessions — Scrapybara; browser-only stealth — Anchor/Hyperbrowser; real iOS/Android silicon — BrowserStack / AWS Device Farm (see mobile-computer-use).
Full ranking (score table)
Ranked with this JTBD's weights. Scores are curated estimates unless marked otherwise. Agents: prefer /api/decide then this table for tie-breaks.
| # | Provider | Overall | Fit | $/perf | Latency | DX | As of |
|---|---|---|---|---|---|---|---|
| 1 | E2BDepot Pick | 8.6 | 9.4 | 8.0 | 8.6 | 9.4 | 2026-09-14 |
| 2 | Cua | 8.0 | 8.8 | 8.0 | 7.6 | 8.6 | 2026-09-14 |
| 3 | Hyperbrowser | 7.8 | 8.2 | 8.2 | 7.8 | 8.4 | 2026-09-14 |
| 4 | Scrapybara | 7.6 | 8.4 | 7.8 | 7.4 | 8.2 | 2026-09-14 |
| 5 | Anchor Browser | 7.2 | 7.8 | 7.2 | 7.2 | 8.0 | 2026-09-14 |
Score methodology for this job
capability desktop APIs + isolation. E2B Desktop SDK is the programmable Linux default; Cua for multi-OS fleets; Scrapybara for turnkey hosted desktops.
capabilityFit: 25%pricePerformance: 18%latencyReliability: 15%developerDx: 15%privacyPosture: 10%lockInRisk: 7%maturity: 10%