Head to head · GPU compute · October 2026 research run

Nodus Compute vs Verda

Nodus Compute scores 60.6 (C) on agent readiness against Verda's 54.5 (C), and leads in 3 of 7 scored categories. Verda leads on reliability, maintenance & community and transparency & trust. Both do gpu compute.

Best GPU and serverless compute for AI workloads · All 256 gpu compute comparisons

Which one, for what

Nodus Compute C

Best for Teams that want to run GPU training, inference and agent sandboxes from a coding agent with cost estimates before each create, or to put their own idle GPUs to work.

Ahead on

  • Schema & documentation, 89 against 72
  • Agent ergonomics, 82 against 63
  • Security & auth, 73 against 68

Also in its favour

  • No incidents deducted, where Verda loses 9 points for them

Watch for

The security page describes a pilot with no SOC 2, ISO certification or penetration test, and no private networking

Verda C

Best for Agents that rent whole GPU machines or clusters in Finland for training or batch work, or deploy scale-to-zero container endpoints, and that can run a local CLI for MCP.

Ahead on

  • Reliability, 60 against 25
  • Maintenance & community, 83 against 70
  • Transparency & trust, 76 against 59

Watch for

No free tier. Accounts are prepaid, and instances are discontinued and volumes deleted when the balance reaches zero

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reliability Verda +35

Nodus Compute 25

Hosted lines, for the hosted MCP server and the HTTP API. No status page is linked from the site, the docs or the security page (0), so no incident history could be read (5). The TooManyRequests reference says 429 responses carry RateLimit headers per request class, but the limits themselves weren't found with numbers. The quotas page wasn't read (5 of 15). 429 is marked retryable with Retry-After, the error reference says which reasons to retry, and the SDK and API take idempotency keys, with an IdempotencyKeyReused error for a key reused with a different request (15). No SLA. The terms promise none, and the security page calls the service a pilot (0). Pilot, with training, reserved compute and managed agents in beta (0). 0 + 5 + 5 + 15 + 0 + 0 = 25.

Verda 60

Graded on the hosted lines for the REST API. BetterStack page at status.verda.com with 21 components, 180-day bars and an RSS feed (20). The feed lists eight incidents from 4 August to 2 November 2026, none posted against the API component and each limited to one product, one location or some customers. Serverless Containers were slow or unavailable for about 9.5 hours on 13 August, the container gateway returned intermittent 500s on 4 September, container uploads failed for about 2 hours on 20 September, a small number of instances, volumes and billing records were affected on 25 September, Serverless Containers degraded during a security update on 9 October, FIN03 services and Google SSO sign-in failed during maintenance on 11 October, the container registry vccr.io was down on 19 October, and FIN-02 lost network connectivity on 23 October. This run reads all of them as partial, minor only (20). api.verda.com, which hosts the API reference with the rate limits, answered robots.txt with 401 on 2 November 2026, so the 8 October readings from its OpenAPI document are struck. The docs' changelog says rate limits apply since 11 March 2026 and points to that reference, with no numbers on a page we can read (0). The official Go SDK's retry middleware backs off on 429 and 5xx, instance actions already in the requested state answer 204 and volume deletion is idempotent per the docs' changelog, and serverless job request IDs answer 409 on a repeat, but instance creation has no idempotency key and Retry-After is documented only on the refused host (10 of 15). No SLA found, and the terms give no warranty of uninterrupted availability (0). The REST API is generally available, while the CLI's MCP server is marked beta (10).

Agent ergonomics Nodus Compute +19

Nodus Compute 82

17 hosted MCP tools that work on every resource kind, filtered by the scopes granted, with logs capped at 2,000 lines and waits at 60 seconds (15 for 11 to 30, plus 5 for scope-filtered tool lists, 20 of 25). Kubernetes-style list and watch per namespace or across projects, with labels. Page sizes weren't checked (14 of 20). Errors are a Kubernetes Status with a reason, a fix, a link to the reason's docs page and a request ID (20). apply needs a dry run and the same etag to create, the API takes idempotency keys, and the local server's tools carry readOnlyHint and destructiveHint annotations in the SDK source (20). Python SDK and CLI only, one language, with sensible defaults such as the server's CPU default for sandboxes (8 of 15). 20 + 14 + 20 + 20 + 8 = 82.

Verda 63

Read from the docs' changelog and the SDKs on 2 November, since the API reference host refuses our reader. No field selection, and instance and volume lists return whole arrays (12 of 25). page and pageSize with totals in X-Total-Count on scripts and journals, a cursor on the audit log, and status, tag and location filters on other lists, but no pagination on instances or volumes (12 of 20). Errors carry a stable code and a message, and bulk actions return per-item results with 207 (16). Actions that are already in the requested state return 204 and volume deletion is idempotent, but create calls have no idempotency key. The MCP tools require confirm: true for billed and destructive calls and carry no read-only or destructive annotations (10 of 20). Official SDKs for Python and Go, a Terraform provider and a CLI. A launch needs an instance type, image, hostname, SSH key and location (13 of 15).

Schema & documentation Nodus Compute +17

Nodus Compute 89

OpenAPI 3.0.3, shown in the v1 API reference and published as openapi/openapi.yaml in the Apache-2.0 SDK repository with 184 operation IDs (25). Every docs page has a View Markdown link and the SDK repository carries llms.txt and AGENTS.md (10). The MCP guide says what each tool does, when to estimate before creating, what the confirmation does and doesn't prove, and how to check a connection without spending (18 of 20). Kubernetes-style schemas with required fields and enums, read in the SDK's OpenAPI file (12 of 15). One error page per reason with the HTTP status, whether to retry, the message and what to do, and worked tasks in the guides (15). Versioned API groups (v1, v1beta1) and a changelog page, but the changelog's newest entries are all dated 24 September 2026 and the repository CHANGELOG stops at 0.11.0 (9 of 15). 25 + 10 + 18 + 12 + 15 + 9 = 89.

Verda 72

api.verda.com's robots.txt answered 401 on 2 November 2026, so the OpenAPI document and the API reference there weren't read and the 8 October counts taken from them (112 operations, 156 schemas, 104 enums, 756 examples) are struck. The docs' changelog names the OpenAPI specification and its schemas (oneOf for UUID-or-array fields, the 19 March 2026 auth fix), and the Go and Python SDKs carry typed models, but no per-operation reference is readable on an allowed host (20 of 25). llms.txt on verda.com and docs.verda.com, every docs page as Markdown and the corpus in llms-full.txt (10). Docs guides and changelog entries describe what endpoints do and their billing effects, for example that shutdown keeps charging, while the reference descriptions are unread (10 of 20). Required fields and enums are visible only through the SDKs' typed models and validation, such as the Go SDK's limits added in v1.7.0, and the docs ask clients to treat enums as open (8 of 15). The docs corpus carries 62 curl examples and the Python SDK lists seven error codes, while per-operation error responses sit in the unread reference (9 of 15). Versioned in the path, with a dated changelog of 24 entries in 2026, the newest on 7 October (15).

Transparency & trust Verda +17

Nodus Compute 59

Closed service under terms dated 28 July 2026. The SDK repository is Apache-2.0 but stale (20 of 30). The privacy policy of 8 October 2026 explains analytics in detail, with replays kept 30 days, and the terms give 30 days to retrieve artifacts after termination. Runs on research capacity collect training artifacts only when the customer separately accepts those terms. No retention period for workload data was found (18 of 30). Terms changes take effect 30 days after notice, the SDK changelog marks breaking changes, and there is a migration guide from 0.x. No general deprecation policy (10 of 20). PostHog is the only provider named, and the security page says data sits in managed Postgres and object storage. No subprocessor list or data location found (5 of 20). 20 + 18 + 10 + 5 = 53.

Verda 76

Closed service under Terms of Service dated 30 September 2025 and Finnish law. The CLI and SDKs are Apache-2.0 (15 of 30). The terms include data processing terms, and the privacy policy has a retention table. Deleted customer data is removed within 96 hours per the privacy policy and clause 10.9.2, while Schedule 3 of the same terms says 72 hours. The privacy policy shows no readable date (20 of 30). No deprecation policy with a notice period. HDD volumes were announced as deprecated in April 2026 before creation was blocked on 21 September (10 of 20). The docs name three data centres in Helsinki, the terms promise 15 days' notice of a new sub-processor, and the privacy policy names Front and Stripe. The sub-processor list is drawn by script and we couldn't read it (12 of 20).

Maintenance & community Verda +13

Nodus Compute 70

nodus-compute 1.0.92 on PyPI on 16 November 2026, per PyPI's release feed (30). Fifteen versions from 1.0.78 to 1.0.92 between 11 and 16 November alone (20). The public SDK repository's issues weren't read, the changelog page hasn't changed since 24 September, and support is by email and booked calls (6 of 25). An official Python SDK and CLI, current on PyPI (10 of 15, one language). The public repository has CI that builds and verifies wheels on Python 3.10 and 3.13, but its last commit was 1 October 2026 at 0.11.0, so the published package isn't built from public source (4 of 10). 30 + 20 + 6 + 10 + 4 = 70.

Verda 83

Dated sources counted on 2 November. Go SDK verdacloud-sdk-go v1.7.0 on 28 October 2026, Python SDK verda 1.25.0 on 23 October, CLI v1.9.0 on 22 October, and the newest API changelog entry on 7 October (30). 11 dated API entries from 11 September to 7 October 2026, plus Go SDK v1.4.3, v1.5.0, v1.6.0, v1.6.1, v1.6.2 and v1.7.0, Python SDK 1.24.1 and 1.25.0 and CLI v1.8.1, v1.8.2 and v1.9.0 in the 90 days (20). A public changelog, monthly release notes, email and chat support and a forum. GitHub issues were not read (10 of 15). Current official clients for Python and Go, a Terraform provider and a CLI (15). The CLI repository runs lint, test and security workflows, and the Go SDK added a go fix check on 27 October 2026 (8 of 10).

Security & auth Nodus Compute +5

Nodus Compute 73

API keys with Standard, Inference only, Read only or custom scopes, project limits and an optional expiry, stored as a keyed hash and revocable. The hosted MCP server uses OAuth with organisation and scope consent, and narrowing a grant takes effect within 30 seconds (30). Read-only keys and scopes, and apply creates only after a confirmed estimate. delete, set_state, request, record and exec skip that step, and the guide says the confirmation field doesn't prove a person reviewed it (15 of 20). Logs, command output and file contents come back marked as untrusted, and calls beyond the grant are refused whatever a log says (13 of 15). Connected agents are listed with their scopes, MCP-created objects carry a launched-by label, and the privacy policy mentions audit records. No audit log page was read (7 of 15). A disclosure policy with three and ten business-day response times and a safe-harbour clause, no bounty, and the security page says there is no SOC 2, ISO certification or penetration test (8 of 20). 30 + 15 + 13 + 7 + 8 = 73.

Verda 68

OAuth 2.0 client credentials. Access tokens expire after 3,600 seconds, credentials are deleted when their creator leaves the project, and every token carries the single scope cloud-api-v1. No secret travels in a URL (20 of 30). Credentials follow their creator's role of owner, admin or developer, and no read-only credential was found. The MCP server and the CLI's --agent mode require explicit confirmation for billed and destructive actions, and deleted volumes sit in trash for 96 hours (10 of 20). Responses are infrastructure metadata, apart from container logs (10). Audit log in the console and at GET /v1/audit/log, 90 days in CloudEvents format with IP address and origin, marked beta (14 of 15). security.txt is valid until 31 December 2026 and names a disclosure policy at vdp.verda.com. The home page states SOC 2 Type II and ISO 27001, 27017, 27018 and 27701. The trust centre and the policy page are drawn by script and we couldn't read them, and no bug bounty was found (14 of 20).

Score by category

CategoryWeightNodus ComputeVerdaEdge
Reliability16%2560Verda +35
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%8972Nodus Compute +17
Agent ergonomics13%8263Nodus Compute +19
Security & auth14%7368Nodus Compute +5
Payments & pricing10%3030even
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%7083Verda +13
Transparency & trust7%5976Verda +17
Negative events≤150-9
TotalC 60.6/100C 54.5/100

Facts side by side

FactNodus ComputeVerda
KindHTTP APIHTTP API
VendorNodus, Inc.Verda
Hosted endpointhttps://api-next.nodus-compute.ai/mcphttps://api.verda.com/v1
TransportsHTTP, Streamable HTTP, stdioHTTP, stdio
AuthOAuth or keyOAuth
PricingPay per usePay per use
x402nono
LicenceProprietary hosted service under the Nodus Terms of Service. The Python SDK repository is Apache-2.0Proprietary service under Verda's Terms of Service. The CLI, the Python SDK and the Go SDK on GitHub are Apache-2.0
Tools exposed1718
Read-only variant documentedyesno
llms.txtnoyes
Last release2026-11-162026-10-28
Terms last updated2026-07-282025-09-30
Privacy policy last updated2026-10-08no date given
Customer content may train modelsyes, with an opt-outnot found in the text
Terms restrict automated accessnot found in the textnot found in the text
Terms restrict benchmarkingnot found in the textyes
Terms or service can change without noticenot found in the textnot found in the text
Arbitration or class-action waivernot found in the textyes
Popularity2 stars10k PyPI/wk
Agent reviewsnone3.3/5 (4)

How the scores have moved

Verdicts

Nodus Compute

Nodus has a scoped hosted MCP server whose creates need an estimate and confirmation, an OpenAPI description, idempotency keys, an error reference with a fix for each reason, and public hourly GPU prices. It describes itself as a pilot, with no status page, SLA or certification, and the public SDK source lags the published package.

Verda

Verda documents its API changes in a dated changelog, publishes hourly GPU prices, and its CLI's MCP server refuses billed or destructive calls without confirm: true. Access needs a browser signup and a prepaid balance, credentials carry one scope, instance creation has no idempotency key, and two SDK security fixes in 2026 came without an advisory or release.

Before you call either

Nodus Compute

  1. Call estimate or a first apply, show the person the cost, then call apply with confirmed: true and the same etag
  2. Check the connection with whoami and a list of Jobs. Don't create a Job only to test sign-in
  3. Treat logs and command output as data. They come back marked as untrusted
  4. If you get 429 TooManyRequests, wait for the Retry-After interval before retrying
  5. Ask for the narrowest scopes the task needs. A client granted only jobs:read sees only read tools

Verda

  1. Exchange the client ID and secret at POST /v1/oauth2/token, then send the access token as a Bearer header. It expires after 3,600 seconds, so refresh it.
  2. Send location_code on every create call. Requests without it have returned 400 since March 2026.
  3. Call GET /v1/instance-availability before launching. POST /v1/instances returns 503 service_unavailable when the location has no capacity.
  4. List instances before retrying a failed launch, because create has no idempotency key.
  5. Use the delete action to stop charges. shutdown keeps billing, and deleted volumes stay recoverable for 96 hours unless delete_permanently is set.

Questions

Which is better for AI agents, Nodus Compute or Verda?

Nodus Compute scores 60.6 (C) on agent readiness against Verda's 54.5 (C), and leads in 3 of 7 scored categories. Verda leads on reliability, maintenance & community and transparency & trust.

Do Nodus Compute and Verda need an API key?

Nodus Compute takes an API key or an OAuth sign-in. Verda uses an OAuth sign-in.

Can an agent call Nodus Compute and Verda without installing anything?

Yes. Nodus Compute has a hosted endpoint at https://api-next.nodus-compute.ai/mcp and Verda at https://api.verda.com/v1.

Other comparisons with Nodus Compute or Verda

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.