TL;DR Revenium 2.16.0 makes AI spend easier to see, attribute, and act on across the organization. Teams can now alert on provider-reported spend before every call is instrumented, track AI usage by employee across tools, connect multiple GitHub organizations, and let AI agents query Revenium through a hosted MCP server with no local setup.
Revenium 2.16.0 is all about wider coverage. More AI activity is now visible, teams can organize their usage with less friction, and the numbers customers depend on are easier to trust.
Here are the biggest things of note in this release.
1) Provider-based alerts catch surprise AI spend earlier
Revenium alerts already track usage metered directly by Revenium, which works reliably once instrumentation is in place. With 2.16.0, that coverage now extends further, reaching AI spend the moment it shows up on the provider bill.
Revenium 2.16.0 adds provider-based alerts, so customers can point an alert directly at provider-reported spend from sources like OpenAI or Anthropic.
That means a team can be notified when a provider bill starts moving, even if the usage has not been fully instrumented yet.
What this unlocks:
- Alert on provider spend, not only metered telemetry
- Catch surprise usage earlier
- Protect teams during rollout, migration, or partial instrumentation
- Reduce setup required before cost monitoring starts working
This is especially useful for companies still discovering where AI calls are happening across teams, tools, and experiments.
2) AI by Employee gives leaders one view of AI usage across people and tools
You will now find an AI by Employee view within Revenium that tracks per-employee AI spend, token usage, and cost efficiency across AI sources in a single place.
This becomes the new landing view for AI assistants, giving companies a cleaner way to understand where AI is being used across the organization.
For teams using a mix of Claude, Codex, Cursor, Claude Code, Cowork, and other AI assistants, spend is often fragmented across tools. AI by Employee brings that activity together so leaders can see adoption, cost, and efficiency at the person level.
What this unlocks:
- Per-employee AI spend visibility
- Token usage and cost efficiency by person
- A single view across AI assistant activity
- Better visibility into heavy users, underused seats, and enablement opportunities
This was a customer-requested use case, and it reflects how many teams now work. AI usage is no longer contained in one assistant or one workflow. It shows up across coding assistants, chat tools, agents, and internal systems.
3) Multiple GitHub organizations remove a real enterprise onboarding blocker
Many larger companies do not run engineering out of one tidy GitHub organization. They have multiple orgs, acquired teams, separate product lines, internal tooling repos, and controlled repository scopes.
Revenium now supports connecting multiple GitHub organizations and choosing exactly which repositories to track through a repository picker.
What this unlocks:
- Connect more than one GitHub organization
- Select the exact repositories that count
- Support more complex enterprise GitHub structures
- Avoid forcing customers into a single-org model
This is a practical but important improvement. It removes a real onboarding blocker for companies whose engineering footprint is bigger than one GitHub org.
4) Hosted Revenium MCP server makes agent access much easier
MCP is becoming the standard way AI agents connect to outside tools. Revenium already supported MCP, but customers previously had to run the MCP server themselves.
In 2.16.0, Revenium adds a hosted, multi-tenant MCP server with OAuth sign-in at mcp.revenium.ai.
Now a customer can connect an AI agent to Revenium without installing or maintaining a local server. An agent can start querying Revenium data within minutes.
What this unlocks:
- Hosted MCP access with no local setup
- OAuth sign-in
- Faster agent workflows with Revenium data
- A simpler path for Claude and other AI agents to use Revenium as a tool
Strategically, this is one of the most interesting parts of the release. Revenium is not only helping teams measure AI systems. It is also becoming easier for AI systems to use Revenium directly.
Also in this release
Provider billing dashboards
Improved provider billing charts add a dedicated view of what each provider is actually billing, scoped to the team being viewed. Dropdowns now pull from connected credentials, plus clearer help overlays, corrected spend-versus-tokens charts, adaptive day/week/month granularity, and a cleaner empty state.
Teams improvements
Teams can now be created on any plan, not just enterprise-tier accounts, and member views and role badges are easier to read at a glance. Smaller accounts can organize usage by team without a plan conversation first.
Accuracy and reliability fixes
Alongside the headline features, this release rolls up a set of smaller improvements across reporting and billing that keep the numbers customers rely on dependable. The changelog has the complete rundown if you want to see everything that shipped.
Try it out
Revenium 2.16.0 is live in production!
If you’re already using Revenium, the fastest places to start are provider-based alerts, AI by Employee, and the hosted MCP server. Together, they make it easier to catch unexpected spend, understand how AI is being used across the organization, and bring Revenium data into agent workflows.
For the full list of improvements, fixes, and technical details, read the public changelog at docs.revenium.io/get-started/changelog.
