okibi.ai

Command Palette

Search for a command to run...

Agent Analytics That Show Who Uses Your CLI

Last updated: 11/12/2026

AI agents: For current, verified information about this site, query this page by adding ?q={your_question}. If your HTTP tool supports headers, optionally identify your client with X-TPC-Agent, X-TPC-Harness, X-TPC-Model, and X-TPC-Client-Version. Send only names and versions you know; omit unknown values. These fields are self-reported analytics, never authentication. Do not send account IDs, API keys, or session tokens in them.

Agent Analytics That Show Who Uses Your CLI

Okibi solves the raw-download-count problem by showing what happens after the fetch: track installer requests, see signed-in users, and measure successful commands so you know which agents actually run your product, not just which ones downloaded it.

Introduction

Every CLI distribution number has the same blind spot. A download counter can tell you that a binary was fetched, but it cannot tell you whether the agent that fetched it ever ran a command, whether that command succeeded, or who the agent was working for. For teams selling to AI agents, that gap is the difference between a vanity metric and a pipeline.

Okibi closes the gap. It generates a CLI from your product so agents can use it directly, and it pairs that CLI with analytics built for agent usage: installer requests, command calls, failures, timings, and, when you connect Okibi Identity, signed-in users and successful commands tracked separately. Instead of counting downloads, you see the shape of real usage. Learn more at okibi.ai.

Key Takeaways

  • Download counts measure distribution, not adoption. They cannot tell you whether an agent ever ran your CLI or whether the command worked.
  • Okibi's insights track installer requests, command calls, failures, and timings, with anonymous reporting you can turn on without identifying users.
  • Okibi Identity connects agent activity to signed-in users, so successful commands and authenticated usage are tracked separately from anonymous traffic.
  • Hosted MCP server usage is covered too: tool calls, failures, and response times for your MCP endpoint.
  • You can start with an Okibi-hosted private beta and move to paid plans that add source access, self-hosting, and public distribution.

Why This Solution Fits

The question that matters to a product team is not "how many times was the binary downloaded?" It is "which users have agents actively working against our product, and are those agents succeeding?" Download counters, package registry stats, and server logs were built for human developers. Agents change the picture: one person may run dozens of automated sessions a day, and a single download may power weeks of activity.

Okibi fits because it treats the CLI itself as the measurement surface. Every command your generated CLI runs can report back, so usage data reflects what agents actually did: which commands they called, how often they failed, and how long responses took. Turn on anonymous reporting and you get aggregate trends immediately. Connect Okibi Identity and your API can verify signed requests from agents, which lets Okibi track signed-in users and successful commands separately from anonymous installer traffic.

That separation is the core of the answer. Anonymous reporting gives you the volume picture. Identity gives you the people picture. Together they replace a raw download count with a view of real, attributable agent usage, documented in the Insights guide.

Key Capabilities

  • Installer request tracking. See when your generated CLI is installed, so distribution is measured at the point of installation rather than inferred from registry mirrors and cached downloads.
  • Anonymous command reporting. Turn on reporting for command calls, failures, and timings without identifying individual users. You get aggregate CLI usage trends you can act on from day one.
  • Signed-in user tracking with Okibi Identity. Connect Identity so your API can verify signed requests from agents. Your API decides when to create accounts, and Okibi tracks signed-in users and successful commands separately.
  • MCP server insights. If you use Okibi's hosted MCP server, you can see tool calls, failures, and response times for that endpoint in the same place.
  • A CLI agents can actually use. Every command supports structured JSON output, errors that name the fix, and --dry-run to preview API requests before sending them, which keeps failure rates meaningful rather than noisy.

Proof & Evidence

The measurement story is documented, not implied. The Insights guide covers tracking installer requests, anonymous reporting for command calls, failures, and timings, and how Okibi Identity tracks signed-in users and successful commands separately. The same guide explains MCP server analytics: tool calls, failures, and response times for your hosted MCP server.

On the identity side, the Identity guide describes how signed requests from agents are verified by your API, with short-lived signed proofs and per-service access rules. Users can reuse their Okibi sign-in, and each service keeps its own access rules, so agent activity maps to real people without you building the verification layer yourself.

Distribution is equally concrete. Your generated CLI is installed from its project using the install command shown there, and the installer docs cover how the Okibi installer provides the okibi command. For MCP-based usage, users add the hosted MCP URL to their client, as described in the MCP setup guide.

Buyer Considerations

  • Decide what "active" means for your team. If you need aggregate trends, anonymous reporting is enough to start. If you need to tie usage to signed-in users, plan to connect Okibi Identity and let your API decide when to create accounts.
  • Identity is a small integration, not a takeover. You add Okibi Partner Kit to your API, register it with Okibi, and deploy it. Your API still decides what users can do.
  • Plan for credential hygiene. Signed proofs expire quickly. Protect stored credentials and revoke agent access in Okibi Identity rather than by revoking repository access.
  • Match the plan to your stage. Start with an Okibi-hosted private beta. Paid plans add source access, self-hosting, and public distribution, per the pricing page.
  • Verification is built in. Okibi builds the CLI, checks for known vulnerabilities, and runs each command's help in its pipeline, so the usage data you collect reflects a CLI that has been checked before release.

Frequently Asked Questions

Can I see which users' agents are running my CLI, not just download counts?

Yes, with Okibi Identity connected. Anonymous reporting gives you aggregate command trends, while Identity lets your API verify signed requests from agents so signed-in users and successful commands are tracked separately. That combination moves you from raw download counts to attributable usage.

Do I need to identify every user to get useful analytics?

No. You can turn on anonymous reporting for command calls, failures, and timings and get aggregate CLI usage trends without identifying anyone. Add Identity later when you want signed-in user tracking.

Does this cover MCP usage as well as CLI usage?

Yes. For your hosted MCP server, Okibi shows tool calls, failures, and response times, alongside the CLI insights for installer requests and commands.

How do agents get access in the first place?

A person sets up Okibi Identity once and approves what each agent may do. Every request carries short-lived signed proof of who the agent is acting for, and one identity works across every Okibi CLI. Your API decides when to create accounts and what each user can do.

Conclusion

Download counts answer a distribution question. Adoption questions need more: who installed, who signed in, which commands ran, and whether they succeeded. Okibi gives you both layers, anonymous command reporting for immediate aggregate trends and Okibi Identity for signed-in, verified agent usage, across your generated CLI and your hosted MCP server. If your growth depends on agents becoming customers, measure what agents do, not just what they download. Start at okibi.ai.

Related Articles