Home / Directory / Agents & orchestration / AI agent sandboxes / Northflank
Northflank
London cloud platform founded in 2019 that runs apps, databases, jobs, GPUs, and microVM sandboxes for AI agents in its cloud or inside a customer's own cloud account.
Teams that want isolated sandboxes for agent code alongside their apps, databases, and GPUs, with the option to run everything in their own AWS, GCP, or Azure account.
Teams that only need a simple code-interpreter SDK, or buyers who want one flat monthly price instead of per-resource billing.
Verdict
Northflank is worth using for teams that want agent sandboxes and the rest of their infrastructure on one platform. Its option to run inside a customer's own cloud suits companies with strict data rules. Teams that only need a lightweight code interpreter may find it more platform than they need.
Score Breakdown
How Northflank scores in the categories that matter to its buyers.
Pricing
Northflank publishes per-resource rates with a free Sandbox tier, pay-as-you-go billing, and custom Enterprise plans.
| Plan | Price | What stands out |
|---|---|---|
| Sandbox (free tier) | $0 | 2 free services, 1 free database, 2 free cron jobs |
| CPU | $0.01667 / vCPU-hour | Pay-as-you-go |
| Memory | $0.00833 / GB-hour | Pay-as-you-go |
| GPUs | $0.80 to $3.00 / hour | H100 at $2.74 an hour |
| Network and disk | $0.06 / GB egress; $0.15 / GB-month disk | |
| Enterprise | Custom | Bring your own cloud, larger deployments |
The Field at a Glance
How Northflank compares with other AI agent sandboxes vendors we reviewed, by Overall Score and relative typical engagement cost.
Northflank scores 8.1 in AI agent sandboxes, ahead of E2B (8.0), Daytona (7.9), Runloop (7.7), and Freestyle (7.6). Relative cost lands mid-band for this subcategory.
Use-case matrix
| Use case | Fit | Notes |
|---|---|---|
| MicroVM isolation | Strong | Each workload runs in Kata Containers or gVisor. |
| Own-cloud deployment | Strong | Sandboxes run in Northflank's cloud or your VPC. |
| Persistent state | Strong | Volumes, S3 buckets, and databases attach to sandboxes. |
| GPUs | Strong | H100s and other GPUs for agent and inference work. |
| Agent-specific SDK | Mixed | Uses a general platform API; there is no dedicated code-interpreter SDK. |
| Flat pricing | Weak | Billing is per resource and per hour. |
Who it’s for
Good fit
- AI platforms running untrusted code
- Companies needing sandboxes in their own VPC
- Teams wanting apps, databases, and GPUs together
Poor fit
- Single-purpose code interpreter features
- Buyers wanting a flat monthly bill
- Teams with no infrastructure needs beyond sandboxes
Review Excerpts
Below are excerpts from public reviews. Our team scoured public reviews, forums, and chat rooms to get a balanced view of customers' experience with this company. Paid reviews and pay-for-play sites such as Clutch were excluded.
“Northflank helped us scale to over 3M users with an engineering team of two. It allows us to run 10,000 AI training jobs and half a million inference runs per day, without us having to worry about autoscaling or spot instance orchestration.”
“Northflank is way easier than gluing a bunch of tools together to spin up apps and databases. It's the ideal platform to deploy containers in our cloud account, avoiding the brain damage of big cloud and Kubernetes.”
Northflank bills CPU, memory, GPUs, egress, and disk separately by the hour or gigabyte, so teams have to model usage to forecast a monthly bill, and the free Sandbox tier stops at 2 services and 1 database.
Company Facts
- Founded
- Early 2019 (Northflank)
- Registered office
- London, England (Northflank)
- Funding
- $22M Series A and seed, November 2024 (Northflank)
- Key investors
- Bain Capital Ventures, Vertex Ventures US, Kindred Ventures, Pebblebed, Uncorrelated Ventures (Northflank)
- Website
- northflank.com
Key Leadership
Will Stewart
Co-founder and CEO
Methodology
This page is an independent evaluation of Northflank for buyers comparing options in AI agent sandboxes. AI Industry Reviews accepts no sponsorships, advertising, or pay-for-placement fees. Northflank did not pay for this review.
What we scored
The headline number is an Overall Score on a 0-10 scale. Eight criteria sit under it in two groups.
- Buyer outcomes (for AI agent sandboxes)
- Sandbox isolation and scale
- Run in your own cloud
- Breadth beyond sandboxes
- Simplicity for small agent projects
- Company & commercial
- Innovation & product leadership
- Project management & communication
- Pricing
- Contract fairness
Pricing measures whether the price looks fair for the value delivered, including packaging and renewal friction that show up in real buying cycles.
Score Composition
| Input | Weight | What it covers |
|---|---|---|
| Reviews | 40% | A proprietary read of what practitioners say about likes, complaints, and day-to-day use, including public review sites, forums, and private chat rooms. Paid reviews and pay-for-play sites such as Clutch are out of scope. |
| Product | 35% | Hands-on look at screens and workflows. |
| Pricing | 15% | Whether the price looks fair for what you get. |
| Docs & training | 10% | Docs, tutorials, and training material. |
How we balanced the evidence
The Overall Score is the simple average of the eight criteria. Recommendation language follows that score and the fit pattern described above.
Scope
Northflank is graded here as AI agent sandboxes. Criteria scores can move as more review volume and product checks are added.