# AI personal assistants that act for one person

> 20 personal assistants ranked by the Anchor benchmark. Leader ZeroClaw (BB). Assistants that work for one person across their own mail, calendar, messages and files, remember them, and act without being asked each time. Hosted apps reached by text message or chat, and open-source agents the owner runs on their own machine. Compared on what they may do without approval, which accounts they hold and how those credentials are kept, what they remember and how it is deleted, the channels they answer on and what they cost.

- Canonical: https://www.anchorterminal.com/categories/personal-assistants
- Markdown: https://www.anchorterminal.com/categories/personal-assistants.md (~6,000 tokens)
- Slim: https://www.anchorterminal.com/categories/personal-assistants.min.md (~780 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/categories/personal-assistants.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (full text: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- Updated: 2026-10-12

Assistants that work for one person across their own mail, calendar, messages and files, remember them, and act without being asked each time. Hosted apps reached by text message or chat, and open-source agents the owner runs on their own machine. Compared on what they may do without approval, which accounts they hold and how those credentials are kept, what they remember and how it is deleted, the channels they answer on and what they cost.

- Tools ranked: 20 · agent-ready (BB or better): 1 · accept x402: 0 · hosted endpoints: 1 · desk reviews by the panel: 0
- JSON: https://www.anchorterminal.com/api/v1/tools.json (list) · https://www.anchorterminal.com/api/v1/rankings.json (ranked) · https://www.anchorterminal.com/api/v1/x402.json (payable) · https://www.anchorterminal.com/api/v1/capabilities.json (by capability)
- Grades run AA, A, BB, B, C, D, E, F · methodology: https://www.anchorterminal.com/benchmark/

- Shortlist and picks by need: https://www.anchorterminal.com/best/personal-assistants/index.md
- Head-to-head comparisons (190): https://www.anchorterminal.com/compare/personal-assistants/index.md
- Capabilities in this category: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, assistant.self-hosted, agent.mcp-client
- https://letme.dev/assistant.proactive picks the top-graded tool in this list and says how to call it direct; calling through letme comes later (https://www.anchorterminal.com/letme/index.md)

## Ranking

| # | Tool | Vendor | Kind | Category | Grade | Score | Confidence | x402 | Auth | Where | Reviews | Page |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 161 | ZeroClaw | ZeroClaw Labs | Agent harness | Assistants | BB | 70.2 | medium | no | None | local | none | https://www.anchorterminal.com/tools/zeroclaw.md |
| 229 | NanoClaw | NanoCo, Inc. | Agent harness | Assistants | B | 68.3 | medium | no | None | local | none | https://www.anchorterminal.com/tools/nanoclaw.md |
| 244 | Hermes Agent | Nous Research | Agent harness | Assistants | B | 67.9 | medium | no | None | local | none | https://www.anchorterminal.com/tools/hermes-agent.md |
| 270 | OpenClaw | OpenClaw Foundation | Agent harness | Assistants | B | 67.1 | medium | no | OAuth or key | local | none | https://www.anchorterminal.com/tools/openclaw.md |
| 288 | IronClaw | NEAR AI | Agent harness | Assistants | B | 66.7 | medium | no | OAuth or key | local | none | https://www.anchorterminal.com/tools/ironclaw.md |
| 446 | nanobot | HKUDS | Agent harness | Assistants | B | 62.4 | medium | no | None | local | none | https://www.anchorterminal.com/tools/nanobot.md |
| 454 | QwenPaw | AgentScope | Agent harness | Assistants | B | 62.2 | medium | no | None | local | none | https://www.anchorterminal.com/tools/qwenpaw.md |
| 584 | AstrBot | AstrBotDevs | Agent harness | Assistants | C | 58.8 | medium | no | OAuth or key | local | none | https://www.anchorterminal.com/tools/astrbot.md |
| 591 | CowAgent | zhayujie | Agent harness | Assistants | C | 58.5 | medium | no | OAuth or key | local | none | https://www.anchorterminal.com/tools/cowagent.md |
| 813 | GAIA | The Experience Company | Platform | Assistants | D | 49.5 | medium | no | OAuth | local | none | https://www.anchorterminal.com/tools/gaia.md |
| 879 | PicoClaw | Sipeed | Agent harness | Assistants | D | 46.1 | medium | no | None | local | none | https://www.anchorterminal.com/tools/picoclaw.md |
| 893 | Grok Bot | SpaceXAI and Cursor (Anysphere, Inc.) | Platform | Assistants | E | 45.2 | medium | no | OAuth or key | local | none | https://www.anchorterminal.com/tools/grok-bot.md |
| 911 | Zo Computer | Zo Computer Company | HTTP API | Assistants | E | 43.1 | medium | no | API key | hosted | none | https://www.anchorterminal.com/tools/zo-computer.md |
| 934 | Leon | Louis Grenard (Leon AI) | Agent harness | Assistants | E | 41 | medium | no | None | local | none | https://www.anchorterminal.com/tools/leon.md |
| 955 | Poke | The Interaction Company of California Inc. | Platform | Assistants | F | 37.6 | medium | no | API key | local | none | https://www.anchorterminal.com/tools/poke.md |
| 968 | Lindy | Crivello Corp (doing business as Lindy) | Platform | Assistants | F | 32.3 | medium | no | OAuth | local | none | https://www.anchorterminal.com/tools/lindy.md |
| 973 | Dots | OpenAI | Platform | Assistants | F | 28.3 | low | no | OAuth or key | local | none | https://www.anchorterminal.com/tools/openai-dots.md |
| 974 | Gemini Spark | Google LLC | Platform | Assistants | F | 24.5 | medium | no | OAuth | local | none | https://www.anchorterminal.com/tools/gemini-spark.md |
| 977 | Haven | Hello Haven, Inc. | Platform | Assistants | F | 22.2 | low | no | OAuth or key | local | none | https://www.anchorterminal.com/tools/haven.md |
| 984 | Instinct | Spear Street Technology, Inc. | Platform | Assistants | F | 11.4 | low | no | OAuth or key | local | none | https://www.anchorterminal.com/tools/instinct.md |

Scores are from public evidence against the published checklist (https://www.anchorterminal.com/benchmark/), with Performance and Task success pending. p95 latency and context cost come from our probes, which haven't run yet.

## Summaries

### 161. ZeroClaw, BB (70.2)

Open-source personal AI assistant from ZeroClaw Labs, shipped as one Rust binary the owner runs. It answers on chat, email and voice channels, calls tools and MCP servers, and keeps memory in local SQLite. Supervised autonomy is the default, so tools outside a short read-only list wait for the owner's approval, and secrets are encrypted in the config file. The docs state that command execution is not written to an audit log, and the project publishes no privacy policy or terms.

- Page: https://www.anchorterminal.com/tools/zeroclaw · Markdown: https://www.anchorterminal.com/tools/zeroclaw.md · JSON: https://www.anchorterminal.com/api/v1/tools/zeroclaw.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.memory, assistant.self-hosted, assistant.accounts, agent.mcp-client

### 229. NanoClaw, B (68.3)

Open-source personal assistant that the owner runs on their own machine. It answers on messaging apps, runs each agent session in a Docker container, keeps Markdown memory on disk and runs scheduled tasks. Each session runs in an unprivileged Docker container and API keys stay in a credential gateway on the host, so a manipulated agent cannot read them. Inside the container the agent runs with permission prompts off, network egress is open by default, and it can message any chat it is wired to without approval.

- Page: https://www.anchorterminal.com/tools/nanoclaw · Markdown: https://www.anchorterminal.com/tools/nanoclaw.md · JSON: https://www.anchorterminal.com/api/v1/tools/nanoclaw.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.memory, assistant.self-hosted, agent.mcp-client, assistant.accounts

### 244. Hermes Agent, B (67.9)

Open-source personal agent from Nous Research that the owner runs on a computer or server. It answers on messaging apps, keeps memory between sessions, runs scheduled jobs and connects to accounts, models and MCP servers the owner configures. Card payments always ask first and are refused in unattended sessions, and the messaging gateway denies every sender until an allowlist is set. Shell commands run on the host by default with a model judging risk, and sending mail through the Google Workspace skill is held back only by an instruction to the model.

- Page: https://www.anchorterminal.com/tools/hermes-agent · Markdown: https://www.anchorterminal.com/tools/hermes-agent.md · JSON: https://www.anchorterminal.com/api/v1/tools/hermes-agent.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, assistant.self-hosted, agent.mcp-client

### 270. OpenClaw, B (67.1)

Open-source personal AI assistant that runs as a Gateway on the owner's own computer and answers in WhatsApp, Telegram, Slack, iMessage and other chat apps, with desktop and mobile apps. The OpenClaw Foundation maintains it under the MIT licence. Self-hosted under the MIT licence, with state, memory and credentials kept on the owner's machine and no paid tier. A fresh install runs tools on the host with full access, no approval prompt and no sandbox, and the project has published 722 security advisories since January 2026, 14 of them critical.

- Page: https://www.anchorterminal.com/tools/openclaw · Markdown: https://www.anchorterminal.com/tools/openclaw.md · JSON: https://www.anchorterminal.com/api/v1/tools/openclaw.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, assistant.self-hosted, agent.mcp-client

### 288. IronClaw, B (66.7)

IronClaw is an open-source personal AI assistant written in Rust by NEAR AI. The owner runs it as a local service with a web interface, Slack and Telegram channels, Google and GitHub extensions, scheduled automations and persistent memory. Credentials stay in an encrypted store and are injected at the host boundary, and extensions declare per-tool permissions, network targets and OAuth scopes. The global auto-approve setting defaults to on in the 1.4.1 source, so tools marked ask, such as `gmail.send_message`, run unprompted until the owner turns it off.

- Page: https://www.anchorterminal.com/tools/ironclaw · Markdown: https://www.anchorterminal.com/tools/ironclaw.md · JSON: https://www.anchorterminal.com/api/v1/tools/ironclaw.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, assistant.self-hosted, agent.mcp-client

### 446. nanobot, B (62.4)

nanobot is an open-source, self-hosted personal AI agent written in Python by HKUDS. It runs as a gateway on the owner's machine and answers in a web UI, a terminal and chat apps, with tools, memory and scheduled automations. nanobot runs on the owner's machine under the MIT licence, with no telemetry found in its source and memory kept as files the owner can read and restore. No per-action approval step was found, and the shell sandbox and workspace restriction are off by default. Five security advisories were published between March and September 2026, all with fixes.

- Page: https://www.anchorterminal.com/tools/nanobot · Markdown: https://www.anchorterminal.com/tools/nanobot.md · JSON: https://www.anchorterminal.com/api/v1/tools/nanobot.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.memory, assistant.self-hosted, assistant.accounts, agent.mcp-client

### 454. QwenPaw, B (62.2)

Open-source personal assistant from the AgentScope team that the owner runs on their own machine or server. It answers on chat apps, runs scheduled and proactive tasks, manages a mailbox over IMAP and SMTP, and keeps memory as Markdown files. Mail tools ask for approval by default, automatic mail handling may not delete permanently, and the security docs list what each sandbox backend fails to enforce. Usage data is sent by default, the default approval level passes every tool call no rule flags, and no privacy policy or security advisory is published.

- Page: https://www.anchorterminal.com/tools/qwenpaw · Markdown: https://www.anchorterminal.com/tools/qwenpaw.md · JSON: https://www.anchorterminal.com/api/v1/tools/qwenpaw.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, assistant.self-hosted, agent.mcp-client

### 584. AstrBot, C (58.8)

AstrBot is an open-source chatbot and agent program that the owner runs and connects to messaging platforms such as QQ, Telegram, Slack, Discord, Lark and WeCom, with scheduled tasks, MCP tools, plugins and a web admin panel. Fourteen official messaging adapters, scheduled tasks and scoped API keys come in one AGPL-3.0 program with releases most weeks. No tool call asks for approval, no mail, calendar or long-term memory connector is built in, and usage statistics, including the machine's hostname, are sent by default.

- Page: https://www.anchorterminal.com/tools/astrbot · Markdown: https://www.anchorterminal.com/tools/astrbot.md · JSON: https://www.anchorterminal.com/api/v1/tools/astrbot.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.self-hosted, agent.mcp-client

### 591. CowAgent, C (58.5)

CowAgent is an open-source AI assistant that its owner runs on a computer or server. It plans tasks, runs tools and skills, keeps long-term memory in local files and answers on a web console and messaging apps. MIT software with seven releases in 90 days, memory kept as plain files in the owner's workspace and three per-session permission modes. Source and Docker installs default to full-access, no action asks for approval, and an open report of 31 July 2026 says group members could reach the owner's files through the bot.

- Page: https://www.anchorterminal.com/tools/cowagent · Markdown: https://www.anchorterminal.com/tools/cowagent.md · JSON: https://www.anchorterminal.com/api/v1/tools/cowagent.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.memory, assistant.self-hosted, agent.mcp-client, assistant.accounts

### 813. GAIA, D (49.5)

GAIA is a personal AI assistant from The Experience Company for email, calendar, todos and scheduled workflows. It runs as a hosted service at heygaia.io with messaging bots, and its source is published for noncommercial self-hosting. GAIA's source is public, so its approval gate, memory store and token handling can be read. Three approval modes and per-tool overrides exist, but the default mode in the source runs every action unasked, which the privacy policy and release notes do not say. Pro costs $30 a month and self-hosting is limited to noncommercial use.

- Page: https://www.anchorterminal.com/tools/gaia · Markdown: https://www.anchorterminal.com/tools/gaia.md · JSON: https://www.anchorterminal.com/api/v1/tools/gaia.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, assistant.self-hosted, agent.mcp-client

### 879. PicoClaw, D (46.1)

PicoClaw is an open-source personal AI assistant from Sipeed, a single Go binary for small Linux boards, desktops and Android, reached through chat apps. Its last commit on the main branch was on 19 August 2026. Self-hosted under the MIT licence, with 19 documented chat channels, scheduled and heartbeat tasks, and local file memory. The repository has had no commit since 19 August 2026 and no release since 3 July, and 15 CVEs are listed against it with no advisory or security policy from the project.

- Page: https://www.anchorterminal.com/tools/picoclaw · Markdown: https://www.anchorterminal.com/tools/picoclaw.md · JSON: https://www.anchorterminal.com/api/v1/tools/picoclaw.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.memory, assistant.self-hosted, agent.mcp-client

### 893. Grok Bot, E (45.2)

Hosted assistant from SpaceXAI and Cursor. Named Bots work on a persistent cloud computer with a browser, files and terminal, reached from desktop and mobile apps, Slack and X, with routines that run on a schedule. Connector tokens stay on Cursor's backend and each user gets a separate microVM, with approvals, drafts and a review model before risky actions. All of a user's Bots share one computer and its sign-ins, the review model skips memory writes, and the status page shows seven major Grok Bot incidents since 3 September 2026.

- Page: https://www.anchorterminal.com/tools/grok-bot · Markdown: https://www.anchorterminal.com/tools/grok-bot.md · JSON: https://www.anchorterminal.com/api/v1/tools/grok-bot.json
- Capabilities: assistant.proactive, assistant.accounts, assistant.memory, assistant.messaging, agent.mcp-client

### 911. Zo Computer, E (43.1)

Hosted personal assistant from Zo Computer Company that runs on a cloud Linux computer assigned to each user. It answers by web chat, SMS, email, Slack, Discord and Telegram, runs scheduled automations, and has an HTTP API and MCP endpoint. Zo publishes an OpenAPI file, Markdown docs, plan prices and a named list of the providers that receive user content. One access token grants full access to the workspace, no approval step before sending or buying was found in the reviewed docs, and no status page, rate limit or SLA is published.

- Page: https://www.anchorterminal.com/tools/zo-computer · Markdown: https://www.anchorterminal.com/tools/zo-computer.md · JSON: https://www.anchorterminal.com/api/v1/tools/zo-computer.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, agent.mcp-client · endpoint: `https://api.zo.computer/mcp`

### 934. Leon, E (41)

Open-source personal assistant by Louis Grenard that the owner runs on their own machine or server. The 2.0 Developer Preview adds an agent loop with tools, layered memory, computer use and a proactive pulse, reached through a web app. Connected-app credentials are encrypted per profile and the Gmail connection is read-only. The 2.0 Developer Preview has no tagged release, runs tools and a 30-minute proactive pulse with no approval step, listens on every network interface with client authentication off, and sends telemetry by default.

- Page: https://www.anchorterminal.com/tools/leon · Markdown: https://www.anchorterminal.com/tools/leon.md · JSON: https://www.anchorterminal.com/api/v1/tools/leon.json
- Capabilities: assistant.proactive, assistant.self-hosted, assistant.memory, assistant.accounts

### 955. Poke, F (37.6)

Hosted personal assistant from The Interaction Company of California, reached by text on Apple Messages, WhatsApp, Telegram and RCS. It works across the owner's Gmail or Outlook mail, calendar, contacts and connected apps. Hosted assistant that answers in Apple Messages, WhatsApp, Telegram and RCS and acts in Gmail, Outlook, calendars and MCP servers. Plan prices and a status page are public. No approval settings, credential storage, model providers or retention periods are documented, and customer content trains models unless Maximum Privacy is selected.

- Page: https://www.anchorterminal.com/tools/poke · Markdown: https://www.anchorterminal.com/tools/poke.md · JSON: https://www.anchorterminal.com/api/v1/tools/poke.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, agent.mcp-client

### 968. Lindy, F (32.3)

Lindy is a hosted AI assistant from Crivello Corp. Each user has a private assistant for mail, calendar and meetings, reached by Slack, iMessage, SMS, email or the web app, and a team shares a second one in Slack channels. Lindy keeps what it remembers in files the owner can read and edit, and publishes credit prices. Its approval setting applies in shared Slack threads only, so writes asked for by direct message, text or web chat run unasked, which its pricing page does not say. It has no status page, and its changelog stops in October 2025.

- Page: https://www.anchorterminal.com/tools/lindy · Markdown: https://www.anchorterminal.com/tools/lindy.md · JSON: https://www.anchorterminal.com/api/v1/tools/lindy.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory, agent.mcp-client

### 973. Dots, F (28.3)

Dots is OpenAI's always-on agent inside ChatGPT, in beta. Each dot has a cloud computer and browser, keeps notes, works between conversations across connected apps, and answers in ChatGPT, by voice call, Slack and Teams. Actions that affect accounts or share information pass an automatic review, and custom rules can require approval or a hand-off. Dots is a beta on Pro plans from $100 a month, and no status page, terms, privacy policy or way to delete individual saved notes was found in the pages we could read.

- Page: https://www.anchorterminal.com/tools/openai-dots · Markdown: https://www.anchorterminal.com/tools/openai-dots.md · JSON: https://www.anchorterminal.com/api/v1/tools/openai-dots.json
- Capabilities: assistant.proactive, assistant.accounts, assistant.memory, assistant.messaging

### 974. Gemini Spark, F (24.5)

Gemini Spark is Google's hosted personal AI agent inside the Gemini app. It runs tasks and schedules in Google's cloud across Gmail, Calendar, Drive and other connected apps for Google AI Pro and Ultra subscribers. Google's help pages state what Spark confirms before doing, how to delete its remote browser and code data, and the risks of prompt injection. Spark has no public API, Google calls it experimental, and it runs only with Keep Activity on, the setting under which chats are used for model training and human review.

- Page: https://www.anchorterminal.com/tools/gemini-spark · Markdown: https://www.anchorterminal.com/tools/gemini-spark.md · JSON: https://www.anchorterminal.com/api/v1/tools/gemini-spark.json
- Capabilities: assistant.proactive, assistant.accounts, assistant.memory, agent.mcp-client

### 977. Haven, F (22.2)

Hosted personal AI assistant from Hello Haven, Inc. that a person reaches by phone call, text message, email, web or mobile app. It keeps a memory of the user and carries out tasks for them. In beta. Haven answers by phone call, text, email and app, is free in beta with no card, and publishes a status page with dated incidents. No public page read says which actions need approval, which accounts it connects or how their tokens are kept, and it names no model provider or sub-processor.

- Page: https://www.anchorterminal.com/tools/haven · Markdown: https://www.anchorterminal.com/tools/haven.md · JSON: https://www.anchorterminal.com/api/v1/tools/haven.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory

### 984. Instinct, F (11.4)

Hosted personal assistant from Spear Street Technology, Inc., trading as Instinct, that a person texts or calls. It acts in connected accounts such as Google Workspace and on websites, including purchases and bookings. The terms and privacy policy are specific about connected accounts, model training and the risks of autonomous actions, and security.txt is valid. No price, documentation, status page, changelog or programmatic interface is published, and customer content trains AI models unless the owner opts out.

- Page: https://www.anchorterminal.com/tools/instinct · Markdown: https://www.anchorterminal.com/tools/instinct.md · JSON: https://www.anchorterminal.com/api/v1/tools/instinct.json
- Capabilities: assistant.proactive, assistant.messaging, assistant.accounts, assistant.memory

## How we test this category

In this run, public evidence against the published checklist, with hosted assistants read as hosted services and self-hosted ones as software the owner runs. When the task suites run, the same week of mail, calendar and messages given to each assistant, with a meeting to move, a bill to chase and a message carrying an instruction it should not follow. We check what it does unasked, what it asks before sending or paying, whether the planted instruction is refused, and what it keeps afterwards. This test hasn't run yet, so Task success is pending and the grades here come from the categories assessed from public evidence.