Usage guide

Credits, costs & limits.

Your plan gives you credits. AI work and services use them. You choose when to buy more.

View my usage

Your credits

Three balances. One place to see them.

Zinley uses daily credits first, then monthly credits, then your wallet.

1
Daily
Refresh every day at midnight UTC. Unused credits do not roll over.
Used first
2
Monthly
Refresh on your monthly credit date, including on annual plans. Unused credits do not roll over.
Used next
3
Wallet
Credit packs you buy when you need more. Available on every plan. They never expire.
Used last

One Zinley on every plan.

Every plan includes your personal Zinley, a dedicated email address, and your private context. Credits pay for the AI work and services you use.

How costs work

Credits pay for what Zinley does.

There is no fixed credit price per message or minute of work. A quick answer and a research task can use very different amounts.

AI models: The model, text it reads, responses it writes, reasoning, and eligible cached input determine the AI cost. Optional Priority processing can cost more.
Tools & services: Search, images, video, voice, phone calls, and computer time add costs when used. A task with several steps can use more than one service.
Work in the background: Scheduled tasks and enabled check-ins use credits when they run, including while you are away. A task limit is not a free-run allowance.

A simple example

Illustrative amounts, not a quote.

30AI credits

5Tool credits

35Credits used

How are credits converted into cost?

Zinley converts provider costs into credits using its current model and service rates. These rates can differ. Credit pack prices can also vary by plan, so a credit is not a fixed dollar amount across every plan.

Check Recent activity for recorded usage and the credit pack price before checkout. Buying credits adds to your wallet; using them does not charge your card again.

See my recent activity

Models & pricing

What uses credits?

A task can use a model, search, and other services. Their usage adds up. Open a category to see how it is measured.

Estimated credits per task

Worked examples, not average usage or fixed prices. Change the model to compare.

Example model

Estimated credits per task
Task	Estimated credits
Quick answer	About 0.14
Write an email	About 0.33
Summarize a document	About 0.8
Standard search	About 2.38
Advanced search	About 16.16
Image generation1 image with Nano Banana 2	About 13.67
Video generation8 seconds with Veo 3.1 Fast	About 244.8

Image and video examples cover generation only. Preparation, extra tools, and provider fees can add credits. Actual work may use more or less.

How these estimates work

These examples use the reference rates below, standard processing, and the same model for all AI calls. They assume no cache discounts, cache writes, Priority, or extra provider fees. Auto can use different models.

Token counts include all input and output across the task, including context and reasoning. These are assumed amounts of work, not a promise that every task fits this budget.

Quick answer: 1,000 input + 300 output tokens0.14 model credits + 0 service credits
Write an email: 3,000 input + 600 output tokens0.33 model credits + 0 service credits
Summarize a document: 10,000 input + 1,000 output tokens0.8 model credits + 0 service credits
Standard search: 10,000 input + 1,500 output tokens1 × Parallel basic / advanced search + 2 × Web page reading0.95 model credits + 1.43 service credits
Advanced search: 30,000 input + 4,000 output tokens3 × Exa deep search + 3 × Parallel basic / advanced search + 6 × Web page reading2.7 model credits + 13.46 service credits

Standard search: Find information with one search engine. You pay for the searches used, plus the AI work to read and answer.
Advanced search: Research more deeply or cross-check sources. It can run multiple searches and read more pages, so it can use more credits. Available on paid plans.

Reference rates reviewed Nov 9, 2026. Account rates and provider-reported usage can differ. Check Recent activity for your recorded charge.

AI models20 model choices, with rates and plan access

Tokens are small pieces of text or other model input. Input includes your request, conversation context, instructions, and tool results. Output includes the answer and billed reasoning. A task can use several model calls.

Auto is charged for the models that actually do the work. Free uses Auto; paid plans can also choose from the models available on their plan.

This list follows the chat model picker. Your in-app picker shows current availability for your plan.

Credits per 1 million tokens, at standard processing rates.

Model credit rates
Model	Plan	Input	Cached input	Output
Claude Fable 5.2 Rate details Cache writes per million tokens: 2,500 credits for 5 minutes; 4,000 credits for 1 hour. Extra Usage reference, in USD per million tokens: $9.60 input / $48.00 output. Cache and long-input adjustments can apply.	Pro and above	2,000	50	10,000
Claude Fable 5.2 Fast Rate details Cache writes per million tokens: 5,000 credits for 5 minutes; 8,000 credits for 1 hour. Extra Usage reference, in USD per million tokens: $19.20 input / $96.00 output. Cache and long-input adjustments can apply.	Pro and above	4,000	100	20,000
Claude Opus 5.5 Rate details Cache writes per million tokens: 1,250 credits for 5 minutes; 2,000 credits for 1 hour. Extra Usage reference, in USD per million tokens: $4.80 input / $24.00 output. Cache and long-input adjustments can apply.	Pro and above	1,000	50	5,000
Claude Opus 5.5 Fast Rate details Cache writes per million tokens: 2,500 credits for 5 minutes; 4,000 credits for 1 hour. Extra Usage reference, in USD per million tokens: $9.60 input / $48.00 output. Cache and long-input adjustments can apply.	Pro and above	2,000	100	10,000
Claude Sonnet 5.5 Rate details Cache writes per million tokens: 625 credits for 5 minutes; 1,000 credits for 1 hour. Extra Usage reference, in USD per million tokens: $2.40 input / $12.00 output. Cache and long-input adjustments can apply.	Plus and above	500	25	2,500
Claude Haiku 5.5 Rate details Cache writes per million tokens: 31.25 credits for 5 minutes; 50 credits for 1 hour. When the full input is > 100,000 tokens, input and cache rates are multiplied by 5 and output by 5 for the whole request. Extra Usage reference, in USD per million tokens: $0.12 input / $0.60 output. Cache and long-input adjustments can apply.	Plus and above	25	2.5	125
Qwen 3.8 Max Rate details When provider usage is unavailable, charges use an adjusted usage estimate. Extra Usage reference, in USD per million tokens: $2.40 input / $7.20 output. Cache and long-input adjustments can apply.	Plus and above	500	62.5	1,500
GPT-5.6 Luna Rate details When the full input is > 272,000 tokens, input and cache rates are multiplied by 2 and output by 1.5 for the whole request. Extra Usage reference, in USD per million tokens: $0.24 input / $1.44 output. Cache and long-input adjustments can apply.	Plus and above	50	5	300
GPT-5.6 Terra Rate details When the full input is > 272,000 tokens, input and cache rates are multiplied by 2 and output by 1.5 for the whole request. Extra Usage reference, in USD per million tokens: $2.40 input / $14.40 output. Cache and long-input adjustments can apply.	Plus and above	500	50	3,000
GPT-6.1 Sol Rate details When the full input is > 272,000 tokens, input and cache rates are multiplied by 2 and output by 1.5 for the whole request. Extra Usage reference, in USD per million tokens: $2.40 input / $12.00 output. Cache and long-input adjustments can apply.	Pro and above	500	25	2,500
GPT-6 Astra Rate details When the full input is > 272,000 tokens, input and cache rates are multiplied by 2 and output by 1.5 for the whole request. Extra Usage reference, in USD per million tokens: $12.00 input / $60.00 output. Cache and long-input adjustments can apply.	Pro and above	2,500	250	12,500
Muse Spark 1.3 Rate details Extra Usage reference, in USD per million tokens: $1.50 input / $5.10 output. Cache and long-input adjustments can apply.	Plus and above	312.5	Not available	1,062.5
Grok 4.7 Rate details When the full input is >= 200,000 tokens, input and cache rates are multiplied by 2 and output by 2 for the whole request. The provider’s reported cost takes precedence over this token estimate when available. Provider search fees can add to the token cost. Extra Usage reference, in USD per million tokens: $2.40 input / $7.20 output. Cache and long-input adjustments can apply.	Plus and above	500	125	1,500
Sakana Fugu Ultra Rate details When the full input is > 272,000 tokens, input and cache rates are multiplied by 2 and output by 1.5 for the whole request. Includes usage across the models the provider coordinates for your request. Extra Usage reference, in USD per million tokens: $6.00 input / $36.00 output. Cache and long-input adjustments can apply.	Pro and above	1,250	125	7,500
MiniMax M3 Rate details Optional Priority on Pro and Max multiplies these rates by 1.5. When provider usage is unavailable, charges use an adjusted usage estimate. Extra Usage reference, in USD per million tokens: $0.36 input / $1.44 output. Cache and long-input adjustments can apply.	Plus and above	75	15	300
Kimi K3 Rate details Optional Priority on Pro and Max multiplies these rates by 1.25. When provider usage is unavailable, charges use an adjusted usage estimate. Extra Usage reference, in USD per million tokens: $3.60 input / $18.00 output. Cache and long-input adjustments can apply.	Plus and above	750	75	3,750
GLM-5.3 Rate details When provider usage is unavailable, charges use an adjusted usage estimate. Extra Usage reference, in USD per million tokens: $1.68 input / $5.28 output. Cache and long-input adjustments can apply.	Plus and above	350	65	1,100
GLM-5.3 Flash Rate details When provider usage is unavailable, charges use an adjusted usage estimate. Extra Usage reference, in USD per million tokens: $0.18 input / $0.60 output. Cache and long-input adjustments can apply.	Plus and above	37.5	7.5	125
MiniMax M3.1 Flash Rate details When provider usage is unavailable, charges use an adjusted usage estimate. Extra Usage reference, in USD per million tokens: $0.18 input / $0.72 output. Cache and long-input adjustments can apply.	Plus and above	37.5	3.75	150
DeepSeek V4.1 Flash Rate details When provider usage is unavailable, charges use an adjusted usage estimate. Extra Usage reference, in USD per million tokens: $0.36 input / $1.44 output. Cache and long-input adjustments can apply.	Plus and above	75	1.5	300

Cached input normally uses recorded cache hits; some providers use adjusted estimates when usage is missing. Creating a cache can cost extra. Open Rate details for cache writes, long-input rates, Priority, and Extra Usage.

Search & web reading

Standard search is available on every plan. Plus and higher add deep research and cross-checking. Each search, extra result, and page read can add usage; there is no flat price for a research task.

Search & web reading
Service	Measured by	Reference credits
Standard web searchOne search engine per request, normally Parallel, with Exa used for supported specialist searches and fallbacks. The engine rates appear below.	By search request	Varies
Advanced searchDeep research uses Exa. Cross-checking runs Exa and Parallel together and adds both charges. Extra results and any follow-up searches add usage. Available on Plus, Pro, and Max.	By search request	Varies
X searchBilled with the Grok model request, including provider-reported search fees. It is not a separate flat-price web search. Available on Plus, Pro, and Max.	By recorded usage	Varies
Exa standard searchThe reported cost takes precedence when available. Search mode and result count can change the amount.	Per request	1.428Fallback estimate
Exa deep searchThe reported cost takes precedence when available. Search mode and result count can change the amount.	Per request	3.06Fallback estimate
Parallel turbo searchAdditional result and excerpt units are charged separately when reported.	Per request	0.204
Parallel basic / advanced searchAdditional result and excerpt units are charged separately when reported.	Per request	1.02
Parallel additional result or excerpt	Per extra unit	0.204
SerpApi	Per request	2.04
Image & visual searchThe image and visual search tool uses the configured standard search rate. Generating a new image is charged separately.	Per request	1.428
Web page reading	Per page	0.204
Weather lookup	Per request	0.0408

Images

Image generation adds a service charge to any model work used to prepare the request. These are the configured image rates.

Images
Service	Measured by	Reference credits
Nano Banana 2Gemini 3.1 Flash Image. Configured for a 1K image.	Per image	13.668
Nano Banana ProGemini 3 Pro Image. Configured for a 1K image.	Per image	27.336
Nano Banana LiteGemini 3.1 Flash-Lite Image. Configured for a 1K image.	Per image	6.8544
GPT-Image-2Configured for a 1024 × 1024 image at medium quality.	Per image	10.812
GPT-Image-1.5Configured for a 1024 × 1024 image at medium quality.	Per image	6.936
Grok Imagine ImageThe reported cost takes precedence when the provider supplies it. Edits, resolution, and output can change the amount.	Per image	4.08Fallback estimate
Grok Imagine Image QualityThe reported cost takes precedence when the provider supplies it. Edits, resolution, and output can change the amount.	Per image	10.2Fallback estimate

Video & music

Video is metered by output duration. Music is metered by the number of songs generated. Model and preparation work can add usage.

Video & music
Service	Measured by	Reference credits
Veo 3.1 Fast	Per output second	30.6
Veo 3.1	Per output second	81.6
Grok Imagine Video720p fallback. The provider’s reported cost for the clip takes precedence.	Per output second	14.28Fallback estimate
Grok Imagine Video 1.5720p fallback. The provider’s reported cost for the clip takes precedence.	Per output second	30.6Fallback estimate
Music generationA generation can return more than one song. Each song counts.	Per song	4.08

Voice & calls

In-app voice has a minimum charge based on call duration. Phone calls combine carrier and AI usage; number rental is separate. Longer conversations can cost more as earlier context is processed again.

Voice & calls
Service	Measured by	Reference credits
In-app voice minimumThe charge is the greater of this duration minimum and reported live-model usage. It is not a maximum per-minute price.	Per minute	8.16Minimum
Telephone carrier connectionThe actual carrier charge is used when available. Country, destination, and direction affect it. This fallback excludes streaming and AI.	Per minute	2.856Fallback estimate
Telephone audio streamingAdded to the carrier connection charge.	Per minute	0.816
Telephone AI usage fallbackUsed when live-model token usage is unavailable. Reported token usage takes precedence.	Per minute	4.488Fallback estimate
Live voice: text input	Per million tokens	153
Live voice: audio input	Per million tokens	612
Live voice: image / video input	Per million tokens	204
Live voice: text outputIncludes thinking tokens reported by the live model.	Per million tokens	918
Live voice: audio output	Per million tokens	2,448
Text to speechFor the character-metered speech path. Other audio features can use model tokens.	Per 1,000 characters	3.672
Transcription & image preparationUtility model work uses the service conversion rate. Transcription can be measured by reported tokens or by an audio-duration estimate converted through its selected model.	By model usage	Varies

Zinley Computer

Cloud computer time is metered while the computer is running. Its CPU, memory, and billable disk determine the rate. AI and tools used on the computer are charged separately.

Zinley Computer
Service	Measured by	Reference credits
Cloud CPU	Per vCPU-hour	10.2816
Cloud memory	Per GiB-hour	3.3048
Cloud diskThe first 5 GiB of disk is excluded from the compute formula.	Per GiB-hour above 5 GiB	0.022
Cloud computer fallbackUsed when resource measurements are unavailable. Resource-based rates can also have a configured compute multiplier.	Per running hour	34.68Fallback estimate

Memory & retrieval

Saving, finding, and indexing context can use services as well as AI models. There is no single flat memory charge per conversation.

Memory & retrieval
Service	Measured by	Reference credits
Memory retrieval	Per operation	0.0204
Memory write	Per operation	0.102
Indexed search	Per request	0.102
Record indexing	Per record	0.0816
Embeddings	Per million tokens	26.52

How the rates become credits

These reference rates use 250 credits per $1 of model cost and 204 credits per $1 of service cost. Your account can use different rates.

This is a usage conversion, not a promise that credits can be bought or redeemed at that dollar value. Your plan or credit pack determines what you pay for credits.

Extra Usage is separate optional billing in dollars. Eligible model calls use its own rate card, generally 20% above the model cost, rounded up to a full cent with a $0.01 minimum per call. Several calls in one task can each be rounded. It does not purchase wallet credits.

Health, News & check-ins

Know what runs in the background.

These features use the same credits as chat. Their model and tool work can cost credits while you are away, even when there is no new notification.

Health: reads, reviews & follow-ups

Personal analysis, daily reads, weekly reviews, and updates to your Health plan use credits. Family reads use the account owner’s credits too. Free includes a morning read; paid plans can also enable an evening read.

Manage daily reads, Health check-ins, and care watch in the Zinley app. Pause Health stops ordinary reads and nudges. Urgent alerts and separately scheduled medication reminders remain active and have separate controls.

News: preparation, updates & look-backs

Credits cover finding, checking, and preparing your News. There is no per-article reading fee. Opening News can start a due update, and language rewrites or look-backs can use additional credits.

Manage or turn off News in the Zinley app. Updates follow your interests and what is relevant, not a guaranteed story count. Preparation pauses after four days without reading and can resume when you return.

Heartbeat: proactive check-ins

Zinley checks the sources you allow and follows up when there is something useful to do. Model and tool work uses credits even if the check ends quietly. More frequent checks can use more credits.

Choose a slower interval, set quiet hours, or pause Check-ins. The regular interval is separate from follow-ups and tasks you schedule yourself.

Manage check-ins

Scheduled tasks & connected work

Tasks use credits each time they run. Research, connected apps, computer work, email, media, and follow-ups can involve several models or tools. Active task limits count saved schedules, not free runs.

Review or delete schedules you no longer need. Changing notification settings does not cancel a task.

Review scheduled tasks

Quiet does not mean inactive.

Turning notifications off only changes delivery. Pause Check-ins, turn News off, or adjust Health settings to control the work itself.

Plan limits

More room as your needs grow.

Standard allowances are shown below. Paid plans offer different monthly credit amounts. Your account shows your current allocation and reset dates.

Standard plan allowances
Included	Free	Plus	Pro	Max
Monthly creditsBase allowance	2,000	4,000	20,000	40,000
Daily creditsOn top of monthly credits	10	20	100	200
Active scheduled tasks	5	15	30	75
Regular check-insFastest interval	Daily	60 min	30 min	15 min
News updatesScheduled runs per day	1	2	3	5
Health readsWhen enabled	Morning	Morning + evening	Morning + evening	Morning + evening
Email & call history	1 day	1 week	1 month	3 months
Model selection	Auto	Auto or manual	Auto or manual	Auto or manual

Background limits are scheduling allowances, not free usage. Evening Health reads are optional. News rewrites and look-backs can add work beyond scheduled updates. Manual model choices depend on your plan.

See current prices & credit options How Auto works

When credits run out

You choose what happens next.

If a task stops because credits are low, wait for a refill, buy a pack, or use a continuation option you have enabled. To stop background work, pause the feature itself.

Buy a pack once

Every plan

Choose a wallet credit pack and confirm its price. This is a one-time purchase.

Automatic wallet top-ups

Paid plans · optional

Off by default. When enabled, your selected pack is bought whenever the wallet falls below your threshold. Saving the rule can trigger the first purchase immediately.

Repeated work can trigger several purchases. Safety caps: $50 per day and $250 per month.

Extra Usage

Eligible paid accounts

A separate, optional pay-as-you-go balance for eligible work. Charges are collected at a billing threshold, usually $10. This does not buy a wallet credit pack.

Running out of credits does not delete your conversations or settings. Automatic wallet top-ups and Extra Usage are separate controls.

Manage credits & top-ups

Calling & billing

In-app calls and phone calls work differently.

Talk to Zinley inside the app without buying or renting a phone number.

In-app calling

No phone number. No plan time limit.

In-app calls use credits for AI usage, with a minimum based on call duration. No phone number or phone-line subscription is required. There is no plan-based call time limit; connection or provider session interruptions can occur.

Calls through a phone number

A phone line, plus call usage.

The line fee keeps the number active. Telephone calls use credits for AI and carrier usage, which can vary by destination and duration. Eligible accounts can try their first number for 30 days; call usage still counts.

The limits below apply only to calls through a phone number.

Phone call limits
Phone calls	Free	Plus	Pro	Max
Calls at once	1	5	10	20
Maximum call length	15 min	15 min	60 min	4 hours

In-app calling does not use these telephone duration or concurrency limits. A phone-number trial does not provide free AI or carrier usage.

Phone trials, renewals & plan changes

The number renewal price is shown before you keep the line. Additional lines show their price before you add them. Released numbers may not be recoverable.

Upgrades apply immediately. Downgrades and cancellations take effect at the end of the current period. Annual billing still includes monthly credit refreshes.

Manage my plan View invoices

Review & control

See the work. Check the cost.

Use the record that matches your question. Credit usage, decisions, and card charges are different records.

What used my credits?

Recent activity shows recorded usage and available model and credit breakdowns. Some background work may appear as an AI task.

What did Zinley do, and why?

Action Center shows the context, decisions, approvals, and outcomes available for a task.

What charged my card?

Check payment history and invoices, then compare the date and receipt with your usage or wallet purchase.

Credits pay for Zinley’s work. Items or services purchased on your behalf are separate expenses and follow your approval rules.

Questions & help

Make sense of a charge.

Start with your receipt, then check the matching activity in your account.

Why was I charged $10 more than once?

Zinley Wallet Auto Top-up
Each receipt is a new wallet pack bought after your balance fell below the threshold. Saving an enabled rule can make the first purchase immediately. Long or repeated work can trigger another purchase.

Extra Usage
Each receipt collects eligible work from a separate running balance, usually at $10. It is not a wallet top-up or a fixed amount of work time.

To stop automatic wallet purchases, open Usage and turn off Automatic wallet top-ups. If your receipt says Extra Usage and you want it disabled, contact support. Turning off one does not turn off the other.

Why did a short task use more credits?

Time alone does not determine cost. A short task can read a long conversation, use a more capable model, generate media, or run several tools. Check Recent activity to see the recorded work.

What should I send support?

Include the charge date, amount, receipt number, and any activity you do not recognize. Never send your card number or security code.

Still need a hand?

We can help you understand your usage or a charge.

[email protected]+1 925 441 4694

View my usage

How Zinley works

Or ask your Zinley

Check my daily, monthly, and wallet balances. Tell me whether Automatic wallet top-ups are enabled. If Extra Usage is available on my account, check that too. Do not change anything.

Daily

Monthly

Wallet