# Wabi (Wabi)

Feedback Bench, personal agents, built 2026-10-26, window 2026-08-31 to 2026-10-25. Web page: https://feedbackbench.com/#/agent/wabi

| Measure | Value |
|---|---|
| Rank | 7 of 12 (rank range 7–8) |
| Feedback Score | 45.1 (95% interval 44.3–45.9) |
| Popularity | 0.164 (share of voice 0.47%) |
| Customer love | 0.510 (95% interval 0.504–0.515) |
| Place | Popularity 10th of 12 × Customer love 6th of 12 |
| Top quadrant | no |
| Users | 444 |
| Posts counted | 525 |
| Posts that judge the agent | 104 |
| Criteria ranking high / low | 0 / 0 of 53 |

## The brief

Written by large language models from this page's numbers and the posts it quotes, then checked: every number in the text is in the build. Interpretation, not measurement.

Wabi is new to public view. So far users judge it on everyday use and rarely mention price or setup. Praise outweighs complaints in the areas they talk about most, chatting with it and getting tasks done. Chatting makes up 22.2% of what its users rate, against 14.4% for other agents. Complaints are few. They come mostly from chatting, setup, connections and privacy. No area sits clearly above or below the category norm. Wabi ranks 7th of 12 overall, with a Feedback Score of 45.1. No posts yet rate it against another agent.

### Why people choose it


### Why people leave


### What would move its score

Rule: For each area where an agent's customer love is below the category norm: the Feedback Score and rank it would have if that one area were at the norm (0.5) and everything else stayed the same, computed with the same formula as the ranking. An area's pull follows its weight, so an area many users of the category rate moves the score more than one they rarely mention.

| Area at the category norm | Feedback Score | Change | Position |
|---|---|---|---|
| Getting access and set up | 45.1 | +0.0 | 7th of 12 |
| Trusting it with data and safety | 45.1 | +0.0 | 7th of 12 |
| Paying for the agent | 45.1 | +0.0 | 7th of 12 |
| Connecting accounts, sites and devices | 45.1 | +0.0 | 7th of 12 |

No single area moves Wabi far. Moving setup, privacy, price or connections to the category norm leaves the Feedback Score at 45.1 and the rank at 7th each time. Customer love in each of these areas is already close to the norm, so improving any one of them changes little.

### Six weeks

No users posted about Wabi in the first seven weeks. Wabi opened to all users on 19 October. That week 444 users posted about it, giving it a 2% share of voice and 5th place.

## Top requests

What users ask to add or change, most asked first. 23 user-weeks ask for something. Requests do not change the Feedback Score. Rule: A separate pass by a large language model reads every counted post and extracts what the user asks the agent or its vendor to add or change, with the criteria it maps to and a short normalised wording; it does not touch the labels or the Feedback Score. A large language model groups the wordings within each criterion (the first criterion the request maps to) into themes, and a second pass merges themes that ask for the same change under different criteria; a merged theme sits under the criterion most of its requests map to, and code counts it once. A theme counts distinct user-weeks that ask for it, per agent; across agents, one user-week per agent. Themes asked in fewer than 2 user-weeks, and requests that share no theme, are not shown. Examples: up to 3 posts per theme from different users, without slurs: whole posts of 60 to 450 characters with at least 8 words beyond handles and links, posts of at most 280 characters first, then the newest.

| Rank | Request | Criterion | User-weeks | Posts |
|---|---|---|---|---|
| 1 | Custom MCP server support | [Which services have a connector or integration](https://feedbackbench.com/criteria/connect.coverage.md) | 3 | 4 |
| 2 | Phone verification for more countries | [Account creation, sign-in and identity checks fail](https://feedbackbench.com/criteria/access.sign_in.md) | 2 | 2 |

### 1. Custom MCP server support

- Wabi, 2026-10-20, @wabi (X): “@wabi @ekuyda can you add custom mcp support? my emails are already connected to composio. don’t want to connect everything again. could use composio, pipedream or zapier.” [source](https://twitter.com/174813527/status/2112665753790955854)
- Wabi, 2026-10-20, @wabi (X): “@blasmoros @wabi hey blas! as a user, i'd like to add mcps, and if that's not an option, submit a connector. for example i'd like wabi to have access to @tuistdev mcp.” [source](https://twitter.com/16185220/status/2112566001971892711)

### 2. Phone verification for more countries

- Wabi, 2026-10-21, @wabi (X): “@wabi i wan't to try wabi, but can't get my verification code (kz +7 phone number). can you please help?” [source](https://twitter.com/1824495180696915970/status/2112934743729909884)

## Ranks highest on

No criterion where it sits clearly in the top half.

## Ranks lowest on

No criterion where it sits clearly in the bottom half.

Position on a criterion: its place among the agents with at least 30 rated user-weeks there, ordered by customer love. It ranks high when both its position and its 95% interval put it in the top half, low when both put it in the bottom half; neither otherwise, or when fewer than 4 agents qualify.

## All 53 criteria

Criterion love: 0.5 is the category norm. n: rated user-weeks.

### Paying for the agent: Too few posts (customer love 0.499, n 3)

Users rarely bring up price. It comes up 0.33 times as often as for other agents. The few user-weeks that rate it split one in praise to two in complaint. Customer love is within range of the category norm.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Free tier, trial or required plan to start](https://feedbackbench.com/criteria/billing.entry_plan.md) | Too few posts | 0.502 | 0.500–0.506 | 1 | 1 | 0 |
| [Usage allowance size, exhaustion and reset wait](https://feedbackbench.com/criteria/billing.usage_cap.md) | Too few posts | 0.498 | 0.495–0.500 | 1 | 0 | 1 |
| [Agent spends usage on overhead and costly choices](https://feedbackbench.com/criteria/billing.token_efficiency.md) | Too few posts | 0.498 | 0.495–0.500 | 1 | 0 | 1 |
| [Plan contents, bundles and usage meters unclear](https://feedbackbench.com/criteria/billing.plan_clarity.md) | Too few posts | 0.499 | 0.496–0.500 | 1 | 0 | 1 |
| [Subscription price versus value received](https://feedbackbench.com/criteria/billing.price_level.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Buying, upgrading, cancelling and refunding go wrong](https://feedbackbench.com/criteria/billing.purchase_flow.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |

Most recent posts:

- Complaint, 2026-10-25, @wabi (X): “@maxparnian i’m a visual guy i’m on @wabi ‘s side for now. only when i run out of wabi credit, i go back to muse wabi, you need to make your credit usage clear” [source](https://twitter.com/61566429/status/2114201189780983872)

### Getting access and set up: Too few posts (customer love 0.497, n 5)

Setup comes up less often than for other agents. Complaints outnumber praise four to one, but the counts are small and customer love is within range of the norm. Two user-weeks ask for phone verification in more countries.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Account creation, sign-in and identity checks fail](https://feedbackbench.com/criteria/access.sign_in.md) | Too few posts | 0.497 | 0.493–0.500 | 2 | 0 | 2 |
| [Country availability, waitlists and staged rollout](https://feedbackbench.com/criteria/access.availability.md) | Too few posts | 0.498 | 0.496–0.500 | 1 | 0 | 1 |
| [Apps offered for phone, tablet and desktop](https://feedbackbench.com/criteria/access.platform.md) | Too few posts | 0.498 | 0.495–0.500 | 1 | 0 | 1 |
| [Effort needed to set up and train the agent](https://feedbackbench.com/criteria/access.setup.md) | Too few posts | 0.503 | 0.500–0.508 | 1 | 1 | 0 |

Most recent posts:

- Complaint, 2026-10-23, @wabi (X): “@wabi @plaid please launch an ipad client soon. the experience of using the iphone client on an ipad is currently too poor.” [source](https://twitter.com/1887798124674215936/status/2113430406980829551)

### Connecting accounts, sites and devices: Too few posts (customer love 0.498, n 5)

Users discuss connecting apps and devices less often than for other agents. Complaints slightly outnumber praise. Three user-weeks ask for custom MCP server support. That is 4% of all user-weeks asking for it across agents.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Which services have a connector or integration](https://feedbackbench.com/criteria/connect.coverage.md) | Too few posts | 0.497 | 0.484–0.507 | 4 | 2 | 2 |
| [Linking accounts and staying logged in](https://feedbackbench.com/criteria/connect.account_link.md) | Too few posts | 0.498 | 0.494–0.500 | 1 | 0 | 1 |
| [Websites block the agent with captchas or IP bans](https://feedbackbench.com/criteria/connect.site_blocking.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Access to the user's own phone or computer](https://feedbackbench.com/criteria/connect.local_device.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Moving files between agent, apps and user](https://feedbackbench.com/criteria/connect.file_transfer.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |

Most recent posts:

- Praise, 2026-10-20, @wabi (X): “wow. been trying out all personal agents out there like @muse @instinct but @wabi is so far the best one i have used. i can even connect my telegram and whatsapp and ask it to organise all my contacts and business convo. [link] [link]” [source](https://twitter.com/1289159025163001856/status/2112536422431490089)
- Praise, 2026-10-22, @wabi (X): “been having a lot of fun building lil custom agentic personal finance thingies for me &amp; my gf @wabi x @plaid opens up so many possibilities [link]” [source](https://twitter.com/285559117/status/2113327980017541197)
- Complaint, 2026-10-22, Play Store (Play Store): “it is completely unusable, the email connection always fails, whether it's a google or microsoft email, the key is that you need to connect to the email to use this software.” [source](https://play.google.com/store/apps/details?id=ai.wabi.android.app)

### Talking to the agent: Too few posts (customer love 0.512, n 18)

Users bring up chatting with Wabi 1.54 times as often as for other agents. Praise outnumbers complaints two to one. Customer love is within range of the category norm.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [App screens, layout, rendering, localisation and stability](https://feedbackbench.com/criteria/ask.app_client.md) | Too few posts | 0.516 | 0.500–0.534 | 7 | 5 | 2 |
| [Threads, chats and parallel tasks organisation](https://feedbackbench.com/criteria/ask.conversation.md) | Too few posts | 0.510 | 0.499–0.522 | 5 | 4 | 1 |
| [Response and task completion speed](https://feedbackbench.com/criteria/ask.speed.md) | Too few posts | 0.490 | 0.481–0.498 | 4 | 0 | 4 |
| [Reaching the agent through messaging apps](https://feedbackbench.com/criteria/ask.channel.md) | Too few posts | 0.503 | 0.500–0.509 | 2 | 2 | 0 |
| [Avatar, personality and tone of replies](https://feedbackbench.com/criteria/ask.persona.md) | Too few posts | 0.503 | 0.500–0.508 | 2 | 2 | 0 |
| [Speaking with the agent by voice or call](https://feedbackbench.com/criteria/ask.voice.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |

Most recent posts:

- Praise, 2026-10-23, App Store (App Store): “simple and very functional, beautiful interface, i highly recommend!” [source](https://apps.apple.com/br/app/wabi-ai-assistant/id6747768928)
- Praise, 2026-10-20, X search: wabi.ai, Wabi AI, Wabi app, Wabi 2.0, Wabi assistant (X): “always bullish on what @ekuyda builds. been using wabi and love how each project has its own conversation, with other people able to join in. excited to see the new use cases this opens up for autonomous agents. congrats on wabi 2.0! [link]” [source](https://twitter.com/362514924/status/2112337875153093041)
- Praise, 2026-10-20, Play Store (Play Store): “friendly , helpful, even a sense of irony , 5 stars,, my only worry now , is what this poor app must endure after it's done it's work!!? great app..” [source](https://play.google.com/store/apps/details?id=ai.wabi.android.app)
- Complaint, 2026-10-20, @wabi (X): “@scobleizer @wabi 1. spinning random agents for simple tasks. 2. not trying to understand the context and asking questions but loaders keep loading for minutes.” [source](https://twitter.com/478570959/status/2112410152855036388)
- Complaint, 2026-10-20, App Store (App Store): “the app fails constantly, the apps are low value/garbage/just confuse the experience. adds chaos to/hijacks a group chat. do not recommend.” [source](https://apps.apple.com/us/app/wabi-ai-assistant/id6747768928)

### Getting tasks done: Too few posts (customer love 0.519, n 21)

Getting tasks done is Wabi's largest area, though it takes a smaller share than it does for other agents. Praise outnumbers complaints 19 to 2. Customer love is 0.519, within range of the norm.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Writing code and building apps or websites](https://feedbackbench.com/criteria/act.build.md) | Too few posts | 0.507 | 0.501–0.513 | 5 | 5 | 0 |
| [Email triage, cleanup and drafting replies](https://feedbackbench.com/criteria/act.email.md) | Too few posts | 0.505 | 0.501–0.511 | 4 | 4 | 0 |
| [Booking appointments, reservations, travel and tickets](https://feedbackbench.com/criteria/act.reservations.md) | Too few posts | 0.505 | 0.501–0.511 | 4 | 4 | 0 |
| [Web research, monitoring and summarising information](https://feedbackbench.com/criteria/act.research.md) | Too few posts | 0.505 | 0.501–0.510 | 4 | 4 | 0 |
| [Forms, applications, claims and life admin paperwork](https://feedbackbench.com/criteria/act.admin.md) | Too few posts | 0.504 | 0.500–0.509 | 3 | 3 | 0 |
| [Stops short, or hands the task back](https://feedbackbench.com/criteria/act.stops_short.md) | Too few posts | 0.497 | 0.489–0.505 | 3 | 1 | 2 |
| [Calendar entries, conflicts and schedule changes](https://feedbackbench.com/criteria/act.calendar.md) | Too few posts | 0.503 | 0.500–0.506 | 2 | 2 | 0 |
| [Budgets, transactions, bills and investments](https://feedbackbench.com/criteria/act.finance.md) | Too few posts | 0.503 | 0.500–0.507 | 2 | 2 | 0 |
| [Creating writing, images, video and social posts](https://feedbackbench.com/criteria/act.content.md) | Too few posts | 0.501 | 0.500–0.506 | 1 | 1 | 0 |
| [Wrong facts, hallucinations and careless mistakes](https://feedbackbench.com/criteria/act.accuracy.md) | Too few posts | 0.498 | 0.494–0.500 | 1 | 0 | 1 |
| [Buying products, orders and purchase judgment](https://feedbackbench.com/criteria/act.shopping.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Contacting other people and businesses for you](https://feedbackbench.com/criteria/act.outreach.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Operating a browser or desktop by clicking](https://feedbackbench.com/criteria/act.computer_use.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Says it did something it did not do](https://feedbackbench.com/criteria/act.false_reports.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Following instructions, rules and corrections](https://feedbackbench.com/criteria/act.instructions.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |

Most recent posts:

- Praise, 2026-10-24, @wabi (X): “the world origami days start in 4 hours, but the site is inaccessible due to #php2shell. fortunately, @wabi retrieved the program and advised me to practice! [link]” [source](https://twitter.com/47615797/status/2114090626597724488)
- Praise, 2026-10-23, Play Store (Play Store): “found me moneys i was owed that i did notknow about and help me find work” [source](https://play.google.com/store/apps/details?id=ai.wabi.android.app)
- Praise, 2026-10-21, @wabi (X): “@joonasvirtanen @wabi wabi killed letterboxd for me 😉 even better: it shows me movies from my watch list that are available on the streaming platforms i’m subscribed to [link]” [source](https://twitter.com/61566429/status/2112706014743285971)
- Complaint, 2026-10-22, @wabi (X): “@wabi many tasks are unavailable. i even asked him to create a test wallet on a testnet, but he couldn't. i think it's still quite different from muse and cue.” [source](https://twitter.com/1362230966249349120/status/2113155810390061314)

### Running on its own: Too few posts (customer love 0.519, n 12)

Users discuss Wabi working on its own about as often as they do for other agents. Every rating here is praise, across 12 user-weeks with no complaints.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Unprompted suggestions, insights and check-ins](https://feedbackbench.com/criteria/auto.proactive.md) | Too few posts | 0.513 | 0.506–0.522 | 9 | 9 | 0 |
| [Multiple agents, teams and delegation between them](https://feedbackbench.com/criteria/auto.multi_agent.md) | Too few posts | 0.505 | 0.500–0.510 | 3 | 3 | 0 |
| [Where the agent runs: its own computer and always-on running](https://feedbackbench.com/criteria/auto.cloud_env.md) | Too few posts | 0.496 | 0.488–0.500 | 1 | 0 | 1 |
| [Remembering the user and context across sessions](https://feedbackbench.com/criteria/auto.memory.md) | Too few posts | 0.502 | 0.500–0.506 | 1 | 1 | 0 |
| [Scheduled recurring tasks and daily briefings run](https://feedbackbench.com/criteria/auto.routines.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Reminders and notifications reach the user correctly](https://feedbackbench.com/criteria/auto.reminders.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |

Most recent posts:

- Praise, 2026-10-20, Play Store (Play Store): “wabi is proactive and good on keeping an eye on important things that one may miss out otherwise” [source](https://play.google.com/store/apps/details?id=ai.wabi.android.app)
- Praise, 2026-10-19, X search: wabi.ai, Wabi AI, Wabi app, Wabi 2.0, Wabi assistant (X): “been using wabi 1.0 and now wabi 2.0 for some time. it is fantastic, keeps unearthing stuff i have missed, including some money that might be due to me from an airline i flew in july. keep doing the amazing. [link]” [source](https://twitter.com/2047711670017888256/status/2112227562185486539)
- Praise, 2026-10-19, App Store (App Store): “fully moved to wabi from muse and instinct. wabi has is much more proactive and makes widgets with everything i need in one place. i track my weight, log my workouts and calories in my health thread and lost 13lbs with wabi! best way to organize my life!” [source](https://apps.apple.com/us/app/wabi-ai-assistant/id6747768928)

### Staying in control: Too few posts (customer love 0.504, n 2)

Users rarely discuss control over the agent. It comes up about a third as often as for other agents. Both user-weeks that rate it are praise.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Which AI model runs, how good it is, and choosing it](https://feedbackbench.com/criteria/control.model.md) | Too few posts | 0.502 | 0.500–0.507 | 1 | 1 | 0 |
| [Configuring behaviour, skills and portability](https://feedbackbench.com/criteria/control.customize.md) | Too few posts | 0.502 | 0.500–0.505 | 1 | 1 | 0 |
| [Asks before acting: too little or too much](https://feedbackbench.com/criteria/control.approval.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Seeing what the agent is doing and has done](https://feedbackbench.com/criteria/control.visibility.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |

### Trusting it with data and safety: Too few posts (customer love 0.495, n 3)

Privacy and safety comes up about as often as for other agents. All three user-weeks that rate it are complaints. Customer love here is 0.495, the lowest of any area, but still within range of the norm.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Permission scope requested and isolation between agents](https://feedbackbench.com/criteria/trust.access_scope.md) | Too few posts | 0.495 | 0.488–0.500 | 3 | 0 | 3 |
| [How the vendor collects, keeps and uses data](https://feedbackbench.com/criteria/trust.data_use.md) | Too few posts | 0.498 | 0.495–0.500 | 1 | 0 | 1 |
| [Vulnerabilities, prompt injection and cross-user leaks](https://feedbackbench.com/criteria/trust.security.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |
| [Refuses legitimate requests because of guardrails](https://feedbackbench.com/criteria/trust.refusals.md) | Too few posts | 0.500 | 0.500–0.500 | 0 | 0 | 0 |

Most recent posts:

- Complaint, 2026-10-19, @wabi (X): “damn, these ai agent apps feel a little scaly invasive. like right off the bat @wabi wants to read delete and send emails for me” [source](https://twitter.com/4828981/status/2112222546649387262)
- Complaint, 2026-10-24, Play Store (Play Store): “asked too much details... phone number and email address... scary part is it wants to have full control of my emails. sorry, i'm uninstalling it.” [source](https://play.google.com/store/apps/details?id=ai.wabi.android.app)

### Dealing with the vendor: Too few posts (customer love 0.508, n 4)

Updates and support comes up about half as often as for other agents. Praise and complaints are even, at two each.

| Criterion | Reading | Customer love | 95% interval | n | Praise | Complaint |
|---|---|---|---|---|---|---|
| [Got better or worse over time, and update cadence](https://feedbackbench.com/criteria/vendor.updates.md) | Too few posts | 0.505 | 0.496–0.517 | 3 | 2 | 1 |
| [Support responses and account bans or lockouts](https://feedbackbench.com/criteria/vendor.support.md) | Too few posts | 0.499 | 0.496–0.500 | 1 | 0 | 1 |
| [Outages, errors and agent going unresponsive](https://feedbackbench.com/criteria/vendor.uptime.md) | Too few posts | 0.499 | 0.496–0.500 | 1 | 0 | 1 |

Most recent posts:

- Praise, 2026-10-22, X search: wabi.ai, Wabi AI, Wabi app, Wabi 2.0, Wabi assistant (X): “wabi 2.0! i built my own to-do app 2do with wabi 1.0 and now it’s evolved into an agent i can hand things off to. incredible. i feel in such good hands with this team. way to go, wabi team! 👏 [link]” [source](https://twitter.com/1666578004347351040/status/2113414890945982815)
- Praise, 2026-10-19, X search: wabi.ai, Wabi AI, Wabi app, Wabi 2.0, Wabi assistant (X): “i’ve been using wabi for a while, and this 2.0 update takes it to the next level. personal agents often feel like a solo experience. wabi makes it multiplayer! the interactive mini apps, home-screen widgets and proactive updates put it right at the top of my list. wabi 2.0 is phenomenal! congrats @ekuyda and the @wabi team!” [source](https://twitter.com/62096847/status/2112224610435633168)
- Complaint, 2026-10-19, Play Store (Play Store): “at first, this was a great app. i used it to create several apps of my own. however, after an update, my account was blocked. i was eventually given access again, but all my work had been deleted. i no longer have any trust in the developers. everything you create can be deleted at any time without warning” [source](https://play.google.com/store/apps/details?id=ai.wabi.android.app)
