SiMa.ai Edgematic

sima.ai · Compute · 1 record · JSON

SiMa.ai Edgematic Compute

caveatcardcardphoneDretention-clamped

docs.sima.ai/pages/overview/hello_sima/run_demos.html · cloud VM · compute/sima-ai-edgematic · JSON · compare

Free tier	A free cloud lane onto SiMa.ai MLSoC and MLSoC Modalix silicon: 'Currently Edgematic is in beta and available for free on the Amazon Web Services (AWS) Marketplace. To obtain access, click link. To sign up, you will need an AWS account.' The architecture guide states the point: Edgematic 'provides a cloud-based, low-code environment where users can design, deploy, and benchmark pipelines', with 'No Hardware Requirements: Allows developers to test and optimize pipelines without purchasing a SiMa DevKit'. unit none; metered on wall-clock; free-beta-cloud
Price	none stated
Per-object cap	not applicable
Retention	the beta is the clock and it has no committed end date; each run allocates one device and the job is a GStreamer pipeline, not a general compute job.
Egress	not stated
Identity	card an AWS account: 'Currently Edgematic is in beta and available for free on the Amazon Web Services (AWS) Marketplace... To sign up, you will need an AWS account', so the identity gate belongs to AWS rather than to SiMa. No evidence row states what AWS requires for that account - card, phone number or ID document - or whether any hold is taken: unknown. The record's notes record the card kind as an inference about AWS Marketplace subscriptions, not as a reading.
Limits	retention-clamped the beta is the clock and it has no committed end date; each run allocates one device and the job is a GStreamer pipeline, not a general compute job. source the beta is the clock and it has no committed end date; each run allocates one device and the job is a GStreamer pipeline, not a general compute job.
Binding	retention-clamped the beta is the clock and it has no committed end date; each run allocates one device and the job is a GStreamer pipeline, not a general compute job.
Interface	unknown inferred from the tooling the provider documents; not probed
Tooling	not stated
Composes with	Composes with an AWS account that exists for other reasons: the same account that runs a free-tier instance can subscribe at no stated cost. It breaks at the model boundary, because the pipeline must be a graph of compiled models, so arbitrary training is out of reach.
Reliability	no reliability event recorded
Notes	Not established: the per-account allocation. SiMa publishes no device-minute figure, no concurrency figure and no session length, so the capacity number is null. Alternative explanation tested: that the listing has no live backend, refuted by the device-allocation step in the docs and the live 200, though neither proves a device is idle at any moment. Identity is recorded as card rather than email because the subscription route is an AWS Marketplace offer, and an AWS account normally carries a payment method at signup; that requirement belongs to AWS, not to SiMa, and it was not tested here.
Evidence rung	documented
Checked	2026-10-22 (live); stale on 2026-11-21

Evidence (4)

date	source	what it proved
2026-10-22	docs.sima.ai/pages/overview/hello_sima/run_demos.html	'Currently Edgematic is in beta and available for free on the Amazon Web Services (AWS) Marketplace... To sign up, you will need an AWS account', quoted
2026-10-22	docs.sima.ai/pages/overview/target_architectures.html	'provides a cloud-based, low-code environment where users can design, deploy, and benchmark pipelines' and 'No Hardware Requirements', quoted
2026-10-22	docs.sima.ai/pages/edgematic/introduction.html	the workflow includes 'How to allocate a device for running the pipeline', so the accelerator is cloud-allocated
2026-10-22	aws.amazon.com/marketplace/pp/prodview-a4upqs5h7v4la	broker fetch returned 200 with title 'Palette Edgematic \| AWS Marketplace', so the distribution route is live

More fields (1)

{
 "wave2": {
  "class": "hardware-vendor-programme",
  "free_tier_kind": "free-beta-cloud",
  "shape": {
   "vcpu": null,
   "ram_gib": null,
   "gpu_class": "SiMa.ai MLSoC / MLSoC Modalix edge NPU",
   "gpu_count": null
  }
 }
}

Unknown or withheld values (8)

$.egress.metered egress-unknown
$.egress.terms not-stated
$.interface interface-unknown
$.per_object_cap.notes process
$.per_object_cap.stated not-applicable
$.reliability_note none-observed
$.retention.notes process
$.tooling not-stated