What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Short answer: run /cost inside an API-key Claude Code session to see that session’s token and dollar aggregate. It does not rank individual prompts. To identify expensive prompts, collect transcript-level usage with a release-compatible reporting method, then compare input, output, cache-read, and cache-creation tokens separately.
Claude Code’s official CLI documentation currently describes session continuation, resumption, print-mode output, JSON output in print mode, and verbose turn output—but not a built-in local command that produces a per-prompt token leaderboard. Your account type also changes how usage should be interpreted, so establish authentication and billing context before comparing sessions.
What Claude Code can show, and what it cannot
There are several different ways people refer to “usage,” and they are not interchangeable. A session total, a model-request total, a thread total, a prompt’s input tokens, and an API cost estimate answer different questions.
Recommended Free Tools
| Method | Scope | What it can provide | Important limitation |
|---|---|---|---|
/cost |
Current Claude Code session | Token and dollar usage aggregate for API-key users | It is not a per-prompt ranking |
claude --continue or claude -c |
Latest conversation | Reopens the latest conversation | The CLI reference does not describe it as a token-reporting command |
claude --resume or claude -r |
Selected session by ID | Resumes a specific session | Resumption alone does not rank prompts |
| Print mode, JSON output, and verbose mode | Individual CLI runs or turns | Ways to capture or inspect command output | Do not assume they contain a complete, normalized usage record unless your installed release documents that behavior |
| Managed-agent session viewer | Anthropic’s managed-agent product | Transcript grouped by model request, session usage, and per-thread usage | This product boundary does not establish that local Claude Code sessions have the same viewer |
session-report plugin |
Local transcripts under ~/.claude/projects, according to its listing |
HTML views for tokens, cache activity, subagents, skills, and expensive prompts | Check compatibility, schema assumptions, and what local data it reads before relying on results |
Therefore, the dependable starting point is a session aggregate from /cost. A prompt leaderboard requires a second step: a transcript or report that exposes usage for individual model requests or turns.
#1 Best Overall
- Powerful Turbo Fan:CAGIWIRU Electric Air Duster have reached speed up to 130000RPM,Can efficient remove dust and debris.With LED light and Three Adjustable Speed to meet with different cleaning tasks.
- With Attached USB C Fasting Charging cable.Charing 3 hours,Enjoy up to 240 minutes of use on the lowest setting, For third speed can use for 25mins.with four charging options to meet with your needs.
- Wide Application: attached 5 different nozzles,making it suitable for a variety of scenes, such as car,pet, pc, keyboards, and other electronic devices. It also serves for office and home clean duster.
- Handy Design & Portable:Cordless air duster electric nozzle is Light weight and small size, design and comfortable to hold and space saving, convenient to carry around. You can put it in your room or inside the car without taking up space.
- Economical &Reusable: Built durable materials with a long-lasting battery,our CAGIWIRU air duster it's a sustainable alternative to disposable air cans.
Start by identifying your authentication and billing route
Do this before interpreting any number. API-key usage is metered differently from subscription or enterprise access, and an API list-price calculation is not a subscription quota measurement.
API-key users
For an API-key session, /cost is the documented in-product command for checking token and dollar usage for the current session. Treat the result as an aggregate over that session, not as evidence that one visible instruction consumed a particular number of tokens.
Subscription and enterprise users
Subscription and enterprise access can apply different limits, accounting, or administrative reporting. Do not paste an API price estimate into a subscription-usage spreadsheet and label it “quota used.” Record the account route beside every comparison so that readers of the report know what the figure means.
Record the scope
Write down whether each number covers a whole session, a thread, one model request, or one user turn. A long-running session can contain several threads, subagent calls, tool calls, and model requests. Mixing those scopes makes a ranking look precise while comparing unlike units.
Use /cost for the current-session baseline
- Start or enter the Claude Code session you want to inspect.
- Run
/costin the session interface. - Record the displayed token fields and dollar figure exactly as shown, along with the model, account route, and session identifier if available.
- Label the record “session aggregate.” Do not enter it as a prompt-level value in a leaderboard.
- Repeat the same process for other sessions only when their scopes and metering routes are comparable.
This gives you a reliable baseline for “which session used the most,” provided the sessions are compared under the same account and with the same definition of usage. It does not answer “which prompt inside the session used the most.”
What the command does not tell you
A session aggregate does not identify the individual turn responsible for the total. It also does not, by itself, establish whether a high total came from a long prompt, a large tool result, repeated context, cache creation, cache reads, output generation, or a combination of those factors.
Rank #2
- Screen Cleaner is Your Screen’s New Bestfriend! - EVEO is proud to present a screen cleaner that’s perfectly suited for your TV. Meet your new screen cleaner, computer screen cleaner, laptop cleaner, iPad cleaner, Macbook,tv cleaner or any other electronic device. Our screen cleaner spray and wipe along with the included plush microfiber cloth, easily removes dust, fingerprints & other smudges on the screen’s surface. Clean effectively for a streak-free screen
- Gentle on your Screens - Our screen cleaner spray and wipes is compatible to be used even for the most sensitive LCD, LED, CRT, and OLED screens in proper use. Be at ease knowing that this product is compatible with all the major brands: LG, SONY, Samsung, Sharp, iPad, iPhone, Android screens, Kindle, PC monitors - we’ve got you covered! Can be used as LCD screen cleaner, tv cleaner, smart tv screen cleaner, and more
- Premium Super Soft Plush Microfiber Cloth - Included in this efficient screen cleaning kit. This screen cleaner spray and wipe has our signature plush microfiber cloth specially designed to comply with screens and monitors and leave them streak-free. Has been shown to effectively absorbs stubborn stains, zero streaks, and no fiber shedding. Always use this together with our screen cleaner for the best results
- Easy-to-Use Screen Cleaner Kit + microfiber cleaning cloth - The EVEO Screen Cleaning Kit will easily be your new favorite tool when it comes to taking care of the things you love most, quick and easy solution for a streak-free screen. Simply screw on the sprayer gun on the screen cleaner bottle. Spray the cleaning liquid onto the microfiber cloth, then wipe the screen until desired cleanliness & shine is achieved.
- Quality Screen Care Kit Worth Purchasing - Enjoy a convenient Screen Cleaner anytime, anywhere you need it. Your satisfaction is our top priority. We stand behind the quality of our products and that’s why, you shouldn’t be worrying at all. We manufacture the Best Screen Cleaner. and can be used for laptop cleaning kit, laptop screen cleaner, tv screen cleaner for smart tv. MacBook screen cleaner, monitor cleaner or screen cleaner spray for electronic devices with microfiber cleaning cloth
Why the text you typed is not the full request
Claude Code requests can contain much more than the visible sentence entered at the keyboard. Anthropic’s API usage model includes input and output tokens, prompt-cache creation tokens, and prompt-cache read tokens. Tool definitions, tool-use content, and tool results also contribute to request usage.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesThat is why a short instruction such as “fix the failing test” can occur inside an expensive turn: the model may also receive files, tool descriptions, prior conversation context, command output, and other content. Conversely, a long instruction may be cheap relative to a turn that returns a large output or processes a substantial uncached context.
For a transparent comparison, keep these dimensions visible:
- Input tokens: tokens supplied to the model for the request.
- Output tokens: tokens generated by the model.
- Cache-read tokens: cached context read during the request.
- Cache-creation tokens: context written into the prompt cache.
- Cost: a monetary calculation based on the applicable model rates and metering route.
Do not silently add cache reads and writes to input tokens, or silently omit them. If you publish one “total tokens” column, state the exact sum convention in its heading or accompanying note.
How to build a per-prompt ranking when the CLI has no documented command
The current official CLI reference does not document a local command that ranks Claude Code prompts by token usage. Use a transcript-reporting approach only after checking the documentation and help output for the release installed on your machine.
1. Confirm the capabilities of your installed release
Run claude --help and check the current in-product help. Command names, output fields, transcript formats, and display behavior can change. The documented options for continuing or resuming conversations, print mode, JSON output in print mode, and verbose turn output are useful for managing or capturing runs, but they should not be treated as a token leaderboard unless your release explicitly says they provide one.
Rank #3
- 【Large size】：Measuring 800 x 300 mm/ 31. 5×12 inches, which is wider and taller than others. Large mouse pad can perfectly fit your keyboard and mouse of any size, and there is also ample space for peripherals such as phones, tablets, etc.
- 【LED backlight mouse pad】：Blade Hawks RGB mouse pad provides 7 static modes and 7 dynamic modes will give you the options you are looking for. Power-off memory function remembers the last lighting mode you selected. The super glow fiber Chroma will give you a colorful gaming setup that you want.
- 【Anti-slip rubber base】：Made from Natural rubber, this gaming mousepad will firmly grip your desktop, allowing you to enjoy the ultimate precision in the games you are most passionate about. It also comes with a texture that enhances this ability even further.
- 【Smooth and waterproof surface】: Micro-textured cloth surface, for extremely precise mouse control and a smooth movement. When liquid splashes on the gaming mouse pad, it forms water droplets and slides down. Will not delay your work or gameplay.
- 【Plug and play】: This mouse and keyboard pad is powered by USB, plug and play, no driver required. One button control light modes. Press one time to change the lighting mode, press twice to change the brightness, press and hold about 3 seconds turn ON/OFF.
2. Choose a transcript source
Keep the raw local transcript for each session you intend to compare. The listed session-report plugin says it reads local transcripts from ~/.claude/projects and generates an HTML report with token, cache, subagent, skill, and expensive-prompt views. That listing makes it a plausible convenience route, not a guarantee that every Claude Code release uses the same schema or that every displayed total is correct.
Before installing or running any report, check its compatibility information and review which transcript files it reads. Local transcripts can contain source code, command output, file paths, and other sensitive project data. Keep generated reports private unless you have removed confidential content.
3. Define one ranking unit
Decide whether one row in your analysis means a user turn, a model request, or a complete thread. A user turn may trigger multiple model requests and tools; a model request may include tool results that are not visible in the original prompt. The most defensible ranking uses the smallest unit for which the report supplies complete usage fields and labels that unit clearly.
4. Preserve the raw usage dimensions
For every row, retain the session ID, thread or request identifier, timestamp if available, model, input tokens, output tokens, cache-read tokens, cache-creation tokens, and cost if provided. Add a column identifying whether the row is a session, thread, request, or turn. Do not replace the raw fields with one unexplained total.
5. Sort by the question you actually care about
Use separate rankings rather than one ambiguous list:
- Sort by input tokens to find requests that supplied the largest context.
- Sort by output tokens to find responses that generated the most text.
- Sort by cache-read tokens to find requests that reused the most cached context.
- Sort by cache-creation tokens to find requests that wrote the largest new cached context.
- Sort by cost only when the model rates and account route are recorded with the row.
If you need a broad “processing volume” view, define a formula such as the sum of input, output, cache-read, and cache-creation fields, then label it as an analytical total. It is not automatically the same as billed cost, and it should not be confused with a subscription limit.
Rank #4
- 【Softer and More Comfortable】Vaydeer wrist rest has unique diamond pattern, which is the combination of softness and aesthetics. The materials of wrist rest are improved into higher quality memory foam and covered with silky smooth lycra. The computer wrist rest makes you as comfortable and cushiony as like rest your wrists on clouds.
- 【Ergonomic Wrist Saver】The wrist rests for keyboard and mouse comes with a 17.32×3.15×0.83 inch keyboard wrist pad and a 5.94×3.15×0.83 inch mouse wrist support. Based on ergonomic design, the unique concave shape is the perfect fit for your wrist joints. The wrist rest pad fits most computer keyboards and laptops, improve hand and wrist posture, release your wrist and arm stress.
- 【Non-Slip Rubber Bottom】Featuring an anti-skid silicone base on the bottom, this wrist keyboard support stays firmly in place on your desk, preventing the padding from sliding around, ensuring stable and consistent wrist support during extended computer sessions.
- 【Better Experience & Pain Relief】Our keyboard arm rest is beneficial to alleviate the soreness caused by direct contact and friction between your arm and a hard desk surface, reducing the risk of wrist fatigue or carpal tunnel. The soft texture of memory foam can evenly distribute the pressure around your wrists and provide good support with just enough give.
- 【Helpful in Multiple Scenarios】Whether you're working, studying, writing, typing, gaming, this keyboard and mouse rest combo is an essential accessory to add comfort and support to your hands and wrists. It’s also a great gift for men, women, family, friend, coworker, gamer, teacher, etc.
6. Inspect the largest rows manually
A high row can reflect a tool result, a large file, a repeated context window, cache activity, or a long generated answer. Read the corresponding transcript entry before blaming the user’s typed sentence. Without the session’s usage record, you cannot credibly claim that a specific visible prompt caused a specific token count.
How to compare sessions without producing a misleading leaderboard
Use the same comparison axes for every session. The following fields prevent the most common apples-to-oranges errors.
| Field | Why it belongs in the comparison | Question to answer |
|---|---|---|
| Account route | Metering and limits differ between API-key, subscription, and enterprise access | Are these figures governed by the same billing system? |
| Model | Rates and behavior can differ by model | Were the requests made against the same model? |
| Scope | Session, thread, request, and turn are different units | What exactly does one row represent? |
| Input tokens | Shows supplied context volume | Which session sent the most input? |
| Output tokens | Shows generated volume | Which session produced the most output? |
| Cache-read tokens | Separates reused cached context from ordinary input | Was a large context reused? |
| Cache-creation tokens | Shows context written to the cache | Did the session create a large cached prefix? |
| Cost or rate assumption | Connects usage to money without hiding pricing assumptions | Is this an actual charge, a displayed estimate, or a list-price calculation? |
When one of these values is unavailable, mark it as “not stated” and retain the source report’s scope. Do not fill a missing cache field with zero unless the report explicitly says zero.
Session viewer versus local Claude Code
Anthropic documents a session viewer for its managed-agent product. That viewer can group a transcript by model request and expose session-level input, output, cache-read, cache-creation, and list-cost information, with per-thread usage as well.
This is useful evidence about how managed-agent observability can be structured, but it is not proof that ordinary local Claude Code CLI sessions expose the same official viewer or command. Keep managed-agent reports and local CLI transcript reports in separate categories when documenting your workflow.
Using a listed session-report plugin safely
The official Claude plugin directory lists a plugin named session-report that advertises an HTML report for local project transcripts, including token, cache, subagent, skill, and expensive-prompt views. If you choose this route:
Best Value
- 【No Drill Mounted】Cable management tray can be assembled in as little as 3 minutes or less-simply. Side clips easily clampe the edge of desk, pass through the holes on the sides, and secure excess cables with baskets. Desk cable management supports inward or outward installation to meet your needs in different scenarios, and provides you with great convenience when collecting and organizing wires.
- 【Rubber Pad Protects Desk】Wire organizer under desk built-in soft rubber pads on the clamp protect your desk from scratches, so you will not damage your furniture or office.
- 【Sturdy and Beautiful】Cord management under desk is made of high-quality carbon steel and holds about 15 lbs. The wire basket has 2 holes on each side of the tray to enable your cables to pass freely. Cable baskets are suitable for desk thickness use from 0.4" to 2.4".
- 【Space Saving & Keep Organized】Desk cable management tray holds all your power cords, outlet strips, USB hubs and computer transformers hidden from your feet, and the data cables will not enter the vacuum cleaner to keep your desk clean and tidy. Desk wire organizer is also suitable for use in the kitchen and outdoors.
- 【Reduce Safety Risks】Cable tray under desk keeps all plugs away from your kids, no more worry about tripping. Using 2 holes wire tray saves extra space, helps you easily cabling and wire protection. The mesh design is not easy to accumulate dust, but avoids safety accidents caused by messy wires.
- Verify that the plugin is available for your installed Claude Code release.
- Read its description and permissions before enabling it.
- Confirm which files under
~/.claude/projectsit reads and where it writes reports. - Compare one report’s session total with the value shown by
/costfor the same API-key session, allowing for scope and timing differences. - Investigate discrepancies before using the report for decisions or published metrics.
- Keep reports containing code, prompts, tool output, or paths out of shared locations unless they have been sanitized.
Availability, schema assumptions, and accuracy can vary by release. Treat the plugin as a visualization aid that needs validation, not as a permanent contract for Claude Code’s local transcript format.
Cost, tokens, and subscription limits are different measurements
Token counts describe processing volume. Cost applies model-specific rates and cache pricing. A subscription limit is an account entitlement or usage rule. These values may move in different directions: a request can process many cached tokens while adding relatively little cost, or generate a costly output despite a modest typed prompt.
When reporting spend, record the model, rate assumptions, cache treatment, account route, and date of the calculation. Pricing changes, so do not present an old API list-price estimate as a current charge. When reporting subscription usage, use the plan’s own metering view rather than converting tokens into dollars.
Free tools Windows power users keep installed
One-click scans. No signup required.
Troubleshooting common results
/cost is unavailable
Check how the session is authenticated and consult the current in-product help. The documented command is specifically the starting point for API-key users; another account route may expose different reporting.
The report shows only one total
Do not infer prompt rankings from a session total. Look for a request- or turn-level view, or retain the session as an aggregate and state that individual attribution is unavailable.
Cache columns are missing
Mark cache-read and cache-creation values as not stated. Do not compare that report with one that includes cache fields unless you explain the difference or restrict the comparison to shared fields.
The plugin total does not match /cost
Check whether the plugin covered the same time range, session, threads, models, and usage dimensions. A transcript report may be incomplete, may use a different sum convention, or may have been generated before the session ended.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →A short prompt appears at the top of the ranking
Inspect the request context and tool activity. Tool definitions, tool-use content, tool results, prior messages, and cache operations can dominate the request even when the visible instruction is brief.
A repeatable reporting checklist
- Identify API-key, subscription, or enterprise authentication.
- Capture
/costfor API-key session aggregates. - Check
claude --helpand current in-product help before relying on undocumented output. - Keep session, thread, request, and turn scopes distinct.
- Preserve input, output, cache-read, and cache-creation fields separately.
- Record the model and account route beside every value.
- Use a transcript report only after checking release compatibility and local-data handling.
- Sort by the metric that answers your question instead of inventing one universal “tokens” number.
- Validate large or surprising rows against the underlying transcript and tool activity.
- Separate token volume, API cost estimates, actual charges, and subscription limits.
Bottom line
For an API-key session, /cost is the documented and immediate way to see total token and dollar usage, but it cannot identify the most expensive individual prompt. Finding prompt-level leaders requires a compatible transcript-reporting method, careful separation of input, output, cache-read, and cache-creation tokens, and explicit labeling of session, thread, or request scope. Treat local reports and plugins as release-dependent tools, and never equate a visible prompt’s length with the full model request.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
