Mycroft
Apply for a pilot

Web app in development. Taking five pilot teams.

Where your Claude Code money goes, and what’s worth changing.

Mycroft reads the Claude Code metrics you already send to Grafana Cloud or Datadog with one read-only credential, prices every token with the prompt cache counted, and writes a plain report: where the money went, which changes are worth making, and what the data cannot tell you.

Apply for a pilot See what a report finds

[email protected] Open in Gmail

Reads daily metric totals. Never prompts, responses or code, and it cannot change anything.

Works with
  • Grafana Cloud
  • Datadog
Sample team / Reports / 1 to 28 Aug 2026 Preview

Anthropic usage, 1 to 28 August 2026

$15,365.62at list prices

24 developers4 findings

By model

  • claude-opus-5$13,398.84
  • claude-sonnet-5$1,966.78

By request kind

  • main$11,631.10
  • subagent$1,966.78
  • auxiliary housekeeping$1,767.74

By developer

Four findings. The largest ceiling is at most $5,884.62 over the month. Read the findings

01What a report finds

Four findings on a $15,365.62 month.

Each dollar figure is the most a change could be worth, never a promise. The ceilings overlap, so they are never added up. Every finding ends with what your data cannot tell you.

Sample team / 1 to 28 Aug 2026 / Findings Preview
  1. 139.5% of agent spend went to three developersAttribution, not a lever
  2. 211.5% of spend is the agent’s own housekeeping, on claude-opus-5At most $942.80
  3. 3Moving claude-opus-5 to claude-haiku-4-5 would pay off, but not immediatelyAt most $5,884.62
  4. 426.6% of your bill went on writing the prompt cacheAt most $3,764.64
2

11.5% of spend is the agent’s own housekeeping, on claude-opus-5

Fastest to act on: one settings change
  • $1,767.74 of $15,365.62, 11.5%, went to requests the agent raised for itself or marked as needing low effort.
  • No cache on claude-opus-5 for this traffic, so moving it to claude-haiku-4-5 costs nothing to switch.
At most $942.80over this period. A ceiling, not a saving.

NextCheck what model your agent uses for auxiliary and low-effort requests. If it is claude-opus-5, point it at claude-haiku-4-5 and leave the main path alone.

What your metrics backend cannot tell youWhether your agent lets you choose a model for this traffic separately from the main one.

1

39.5% of agent spend went to three developers

  • 24 developers ran agents in this period, for $15,365.62 in total. The top three account for $6,071.90, 39.5% of everything.

The vertical line is the team median, $464.69.

Attribution, not a leverThis finding says where the money went, not that it went somewhere wrong.

NextAsk [email protected] what they have been running agents on. The useful outcome is usually a technique worth spreading rather than a habit worth stopping.

What your metrics backend cannot tell youWhether any of this is a problem. Spend follows work, and the most expensive developer on a team is frequently the most productive one.

3

Moving claude-opus-5 to claude-haiku-4-5 would pay off, but not immediately

  • claude-opus-5 is 87.2% of spend, $13,398.84.
  • Once each session’s cache is warm on claude-haiku-4-5, the same tokens would cost $5,884.62 less over this period. The first request of every session that moves pays to write its cache again, and daily totals do not count sessions, so that cost is not subtracted here.
At most $5,884.62over this period. A ceiling, not a saving.

NextTry the cheaper model in one repository for a week before changing anything broadly. Routing at session boundaries preserves the cache; switching mid-session destroys it and costs more than it saves.

What your metrics backend cannot tell youWhether the cheaper model would do the work. This compares what the same tokens would cost; it is not a judgement about output quality.

4

26.6% of your bill went on writing the prompt cache

  • 811,852,800 tokens were written to the prompt cache, costing $4,092.00 at the five-minute write rate. At the one-hour rate the same writes would cost $6,547.20.
  • 15,713,280,000 tokens were read back from cache, about 19.4 reads for every token written.
At most $3,764.64over this period. A ceiling, not a saving.

NextLook for what rewrites a warm prefix: sessions restarted or cleared often, tools or MCP servers that change mid-session, model switches inside a session, and long pauses.

What your metrics backend cannot tell youHow many of these writes were a warm prefix written again and how many were new context that had to be written once. Only the first kind can be avoided.

Prices checked against each vendor’s own pricing page on 24 Sep 2026. Did not run: repository concentration. Claude Code does not emit a repository label.

Sample report for an invented team. The arithmetic is real.

02The prompt cache

On Claude Code, the cache money is in the writes.

The cache already works, so “turn caching on” is not the saving. Writing it is what costs: every write is billed above list price, 1.25x for five minutes and 2x for an hour.

On our own Claude Code usage

40.7%

of the bill went on writing the cache, at the five-minute rate, while about 96% of input was read from cache.

One developer over two days, read live from Grafana Cloud, 24 and 25 September 2026.

In the sample team

26.6%

of the bill: $4,092.00 at the five-minute rate, $6,547.20 at the one-hour rate.

Finding 4 above, at most $3,764.64 over the month.

A warm cache gets rewritten by new sessionscompactionsmodel switchestools or MCP servers changing mid-sessionlong pauses

Mycroft shows what share of your bill went on writing the cache, and the most that fewer rewrites could be worth.

The cheaper model isn’t always cheaper.

A model switch is one of those rewrites. On a typical agent step the cheaper model costs 3.9x more on the first request, pays back by the sixth, and is 9.4x worse if you keep switching. Every router and dashboard compares list prices; Mycroft counts the cache before it suggests any model change.

Moving a coding agent from Sonnet 4.6 to Haiku 4.5Cumulative input cost, requests after the switch

Cumulative input cost over ten requests after a model switch Staying on Sonnet 4.6 costs $0.045 per request. Switching to Haiku 4.5 costs $0.177 on the first request, 3.9 times more, then $0.015 per request, so it pulls ahead on request 6 and reaches $0.313 by request 10 against $0.453 for staying. Alternating costs $0.426 every request and reaches $4.260 by request 10, far off this scale.

Switching is ahead by $0.019. It overtook staying on request 6.

One coding-agent step at list prices: about 141,000 cached prompt tokens and 1,000 fresh ones, from published coding-agent workload research. Five-minute cache.

Everybody in this market collects and displays. Mycroft judges.

Named after Mycroft Holmes, Sherlock’s cleverer brother, who never gathered the evidence himself. He worked out what it meant.

03The web app

Preview, in development

A quiet place to read what your bill means.

Not another dashboard: no chart walls, no live tiles. A report, and the few screens around it. Today the same engine runs by hand for pilot teams. These screens are the app we are building around it.

  1. 1Connect

    One read-only credential, which you create, scope and can revoke.

  2. 2What we can see

    The labels your telemetry carries, before any usage is read.

  3. 3The report

    Where the money went, then the findings. Shown above.

  4. 4Since last time

    Next month opens with what moved.

  5. 5Settings

    Revoke the connection, delete stored data.

Seconds after connecting. A gap becomes a stated limit, not a silent error.
Equal windows, compared. It never calls a shorter month a saving.

Sample report for an invented team. The arithmetic is real.

04What it reads

It reads daily totals. Nothing it could change.

Built and tested end to end on real Claude Code data, on both backends, in September 2026.

Reads Queried read-only, as daily totals

  • Token counts by model and type: input, output, cache reads, cache writes.
  • The labels Claude Code attaches: user.email, query_source, effort.
  • Your backend’s own cost figure, to reconcile against.
Grafana Cloud, or any Prometheus-compatible backend

Cloud Access Policy token, metrics:read on one stack. Live-verified, Sep 2026.

Datadog, every site

API key plus an Application key, metrics_read and timeseries_query. Live-verified, Sep 2026.

Never touches Not read, not stored

  • Prompts, responses or code.
  • Files, logs or traces.
  • Anything it could change.

Stored, so next month can be compared: daily totals by model, token type, developer and request label, plus the findings. You can delete them at any time.

Reconciles against your backend’s own number. Where they differ, the report says why. Usually it is one-hour cache writes, which the metrics do not label.

  • 19models priced, from Anthropic, OpenAI and Google, with cache multipliers and minimums
  • 31daysafter the last price check, Mycroft refuses to print a report
  • 1read-only credential, which you can revoke at any time
  • 0prompts, responses, files or logs read

05Roadmap

Where we are.

The engine works and has been checked against live data. The web app is next. The only date here is one that has already happened.

  1. Built

    The engine

    Pricing across 19 models and 3 vendors, cache economics, findings, the monthly comparison.

  2. Live-verified, Sep 2026

    Claude Code via Grafana Cloud and Datadog

    Real telemetry, read end to end on both backends.

  3. Now

    Five pilot teams

    Reports prepared by hand from the same engine, one team at a time.

  4. In development

    The web app

    Connect, the connection check, the report, the monthly loop, settings.

  5. Planned, not built

    More agents and backends

    OpenAI Codex, GitHub Copilot and Cursor, and more metrics backends.

Later, research stage: cache-aware model routing that acts on the findings.

06The pilot

Five pilot teams get a free report on their last two weeks.

Prepared by us, from your own data, in exchange for honest feedback. You shape the web app before it launches.

Apply for a pilot

[email protected] Open in Gmail Outlook

Who it is for: a team that runs Claude Code and sends its telemetry to Grafana Cloud or Datadog, or could.

Not wired up yet? It is a few settings, deployable org-wide through Claude Code’s managed settings, and we know the traps.

How a pilot works

  1. You email us.A line on your team and where your Claude Code telemetry goes.
  2. You create one read-only credential.Grafana: metrics:read on one stack. Datadog: metrics_read and timeseries_query. We agree a secure way to send it; never a plain email or chat thread.
  3. We run the connection check.What your telemetry can and cannot support, and how to fix any gap.
  4. We send the report on your last two weeks.By email. Then revoke the credential; we encourage it.
  5. Optional: a month later, what moved.The second report, comparing equal windows.

07Questions

What platform leads ask first.

What exactly does Mycroft read?

Daily totals of the metrics Claude Code already sends: claude_code.token.usage and claude_code.cost.usage, with their labels for model, token type (input, output, cache read, cache write), developer (user.email), request kind (query_source) and effort. It never sees prompts, responses, code, files, logs or traces.

Can it change anything in our account?

No. It uses one read-only credential that you create and scope yourself. Grafana Cloud: a Cloud Access Policy token with metrics:read on one stack. Datadog: an API key plus an Application key scoped to metrics_read and timeseries_query. You can revoke it at any time, and we encourage you to once your report arrives.

We haven’t sent Claude Code telemetry anywhere yet. Can we still apply?

Yes. It is a few settings, deployable org-wide through Claude Code’s managed settings, and we will help. We know the traps: Grafana needs cumulative counters and silently gets nothing without them. Datadog needs the default delta counters and silently drops cumulative ones.

Why is every figure “at most”, and why not add them up?

A metrics backend cannot know everything about how your agent runs, so each dollar figure is the most a change could be worth over the period: a ceiling, never a promised saving.

The ceilings also overlap. In the sample, the housekeeping traffic is inside the claude-opus-5 total that the model-switch finding prices, so a sum would count it twice.

How do you keep the prices right?

The price table covers 19 models from Anthropic, OpenAI and Google, with cache read and write multipliers, per-model cache minimums and long-context surcharges, rechecked against each vendor’s own pricing page. If it is more than 31 days old, Mycroft refuses to print a report.

Can we use the web app yet?

Not yet. The web app is in development. Today the same engine runs by hand for five pilot teams, and those teams shape the app before it launches. Claude Code is supported today; OpenAI Codex, GitHub Copilot and Cursor are planned, not built.