Per-API-key usage visibility and scoped API key controls for production VFX workflows

[JoeGaffney]

Discussion

We are using ComfyUI Partner Nodes in production VFX pipelines and strongly value the quality of the integration and model ecosystem.

However, from what we can currently see in the usage dashboard and API key system, there are limitations that make it difficult to operate in a production studio environment where cost attribution and access control are required.

I want to raise this as a discussion rather than assume it is not already planned.

It is possible there are roadmap items or existing features I am not aware of, so happy to be corrected if this is already being addressed.

---

Current Observations (based on what is visible today)

• Usage appears aggregated at the account level
• API keys exist but do not have visible per-key cost breakdowns
• No filtering or reporting per API key in the usage dashboard
• No scoped permissions per API key (all keys appear full access)
• Limited export or audit-friendly reporting for finance or production approval workflows

---

What we are trying to solve

In production VFX environments we typically need:

1. Per-API-key cost attribution

• Understand exactly which key generated which cost
• Separate usage across tools, users, or services
• Export per-key usage for internal reporting

2. Basic API key scoping controls

Not full RBAC, but lightweight controls such as:

• Restrict which models a key can access
• Optional rate limits per key
• Ability to separate production vs experimental keys

3. Better audit and reporting clarity

For internal approval workflows we often need:

• Clear per-key usage summaries
• Simple exportable billing breakdowns
• Visibility that can be shared with finance or production stakeholders

---

Why this matters

In studio environments, AI usage needs to be:

• traceable per integration or service
• separable between experimentation and production
• easy to justify for client billing or internal budgeting

Without per-key visibility, we end up having to route everything through a single shared account, which makes governance and approvals harder than necessary.

---

Question / Roadmap clarification

Is there already work planned or in progress for any of the following?

• per-API-key usage breakdowns in the dashboard
• API key scoped permissions (model-level restrictions, limits, etc.)
• improved export or reporting for billing and audit purposes

If so, I may simply not have access to the relevant roadmap information yet and would appreciate being pointed in the right direction.

---

Closing note

We strongly prefer ComfyUI Partner Nodes over alternatives due to the model ecosystem and workflow integration quality.

This feedback is mainly about making it production-viable in environments where cost attribution and access control are required, rather than changing the core offering.

In our current setup, we enforce API key usage via environment-level configuration and override the default authentication flow to ensure consistent key selection across distributed workers.

This is necessary because UI-based authentication is not well suited to headless or distributed production environments common in VFX pipelines.

If this pattern aligns with the intended direction, I would be happy to contribute a PR or help refine an implementation approach.