Facets review

Facets is an infrastructure platform for teams that want AI agents to make changes without giving them an unchecked path to production. Platform teams define contracts describing what infrastructure blocks provide and which rules they must follow. Facets checks requests against those contracts before work runs; some changes can require human approval.

An approved release can coordinate infrastructure, configuration, deployment, routing, monitoring and alerts. Facets says it works with Claude Code, Cursor, Praxis and custom agents, and runs plain Terraform in your own cloud account. It can also connect to existing tools and infrastructure. The homepage says existing resources are not deleted or rebuilt.

That makes Facets less of a coding agent and more of a controlled route from an agent’s request to a cloud change. The pitch is useful, but the key question for a buyer is whether its contracts and approval flow cover the changes their team actually needs to control.

Facets puts the policy check in the release path, rather than relying only on an agent to follow instructions. Contracts let platform teams set rules for infrastructure blocks, and Facets checks each request before changes run. Human approval can be required for some changes. That is a practical shape for agent-led infrastructure work: define what is allowed, then keep a person in the loop where the team wants one.

The release coordination is another useful part of the offer. Facets says it brings infrastructure, configuration, deployment, routing, monitoring and alerts into one release. That could reduce the need to stitch together separate steps for an approved change, though the published materials do not establish how broadly this works across real-world tool stacks.

It also does not require an all-at-once replacement, according to the product description. Facets supports existing tools and infrastructure, and the homepage says it does not delete or rebuild existing resources. It supports several named coding agents as well as custom agents, and runs plain Terraform in the customer’s cloud account. For teams concerned about handing cloud execution to an opaque service, those are relevant design choices. They are not, by themselves, proof that every workflow is covered.

The main obstacle is price. The pricing page lists SaaS at $799 per month, with 20 Resource Instances included and no free plan. There is a 14-day free trial, but that is a short window to assess a platform that sits across infrastructure and release workflows. Cloud-provider charges are separate.

The instance allowance also makes the headline price incomplete for larger setups. The listed rate is $10 per month for each additional Resource Instance through 250, then $8 per month for instances 251–500. Above 500, pricing is custom. Enterprise includes unlimited Resource Instances, self-hosting in your VPC, custom SLA guarantees, 24/7 priority support and a dedicated success manager, but its annual price is negotiated and not published. The SaaS annual rate is not published either.

That leaves buyers with a meaningful cost to evaluate and little public detail in the published materials about outcomes or deployment specifics. Our verdict is promising guardrails for agent-led changes, but expensive to assess and thin on public proof. Teams should get a quote based on their Resource Instance count and verify the contract and approval workflow against their own infrastructure before switching.

Facets’ SaaS plan is $799/month, according to the pricing page, and includes 20 Resource Instances (RIs). There’s no free plan, but the plan includes a 14-day trial. Additional RIs cost $10/month each for the first 250, then $8/month each for RIs 251–500. Above 500, pricing is custom. That makes the advertised $799 a starting point, not a predictable ceiling: cloud infrastructure charges are also billed separately by your cloud provider.

Enterprise pricing is negotiated on a fixed annual basis, with no public rate, so we can’t put a number on the jump from SaaS. It includes unlimited RIs, a self-hosted control plane in your VPC, custom SLA guarantees, 24/7 priority support and a dedicated success manager. The SaaS plan has no published annual rate. Before evaluating, confirm how Facets counts RIs and what the custom price would be at your expected scale; those details can materially change the bill.

Facets says it connects to existing tools and infrastructure, and its homepage says existing resources aren’t deleted or rebuilt. It also runs plain Terraform in your cloud account. That suggests adoption need not start with replacing the underlying infrastructure, but the published materials don’t quantify setup work or spell out a migration process.

The main operational distinction is the control plane: SaaS is Facets-hosted, while Enterprise is self-hosted in your VPC. If you leave, your cloud resources remain in your account, but data export, contract portability and a formal offboarding process are not published in the materials we have. Don’t assume those are automatic; get the export and teardown steps in writing before you commit.

We weren’t given Reddit posts or comments about Facets, so we can’t responsibly say what users there cite as reasons for leaving. Public user feedback is thin, which makes a reference call and written exit terms more useful than relying on unverified churn stories.