# Asmi (slim)

> Asmi is a hosted assistant that calls, texts and emails businesses and people to finish real-world chores, such as booking appointments, cancelling memberships or chasing refunds. It is reached by WhatsApp, the web or an MCP server.

- Full: https://www.anchorterminal.com/tools/asmi.md (~7,000 tokens) · this version ~1,580 tokens · JSON https://www.anchorterminal.com/tools/asmi.json · canonical https://www.anchorterminal.com/tools/asmi
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · Updated: 2026-10-14

**F · 33.3/100 · rank #973 of 1010 · #19 in Personal assistants · not agent-ready · confidence medium**


Assessment: Asmi is a hosted errand service with a 12-tool MCP server for Claude, ChatGPT and Muse, using OAuth with separate scopes for tracking tasks and for contacting people. Tool schemas sit behind sign-in, the outreach scope is ticked by default, and no status page, changelog or security contact was found.

## Facts

- Kind: Platform · vendor: Humint Labs, Inc. · category: Personal assistants · legal entity: Humint Labs, Inc. · provenance 58/100
- Endpoint: `https://prod.asmiai.com/mcp/claude/` (Streamable HTTP)
- Auth: OAuth · pricing: Freemium · x402: no · licence: Closed service under Asmi's Terms of Service (Humint Labs, Inc., a Delaware corporation, effective October 2026).
- Probe metrics: not measured yet (probes haven't run)
- Surface listed: Asmi's errand service and its MCP server. www.asmiai.com is drawn by script, so the home, MCP docs, pricing, terms and privacy text were read from the site's JavaScript bundle on 14 October 2026, with the OAuth protected-resource metadata. No account was created and no MCP call was made.
- Where it runs: A cloud service at prod.asmiai.com. A voice infrastructure provider stores call audio under a data processing agreement and a telephony provider routes calls; neither is named (https://www.asmiai.com/privacy).
- What it acts on: Phone calls, texts and emails to third parties, web search for businesses, task follow-up, and quotes and bookings that the owner approves (https://www.asmiai.com/terms-and-conditions).
- Outside agent access: MCP over HTTP at https://prod.asmiai.com/mcp/claude/, /mcp/openai/ and /mcp/meta/, OAuth with scopes tasks and outreach, or a 60-day bearer key issued by support. Tasks run in the background and are polled with execute_status (https://www.asmiai.com/docs/mcp).
- Company: Humint Labs, Inc., a Delaware corporation. Disputes go to Delaware courts after 30 days of informal resolution (https://www.asmiai.com/terms-and-conditions).
- Eligibility: 18 and older (https://www.asmiai.com/terms-and-conditions).
- Prices: Asmi, 20 tasks $10 per month (plan); Asmi, 100 tasks $49 per month (plan)

- Scores: Reliability 30, Performance pending, Schema & documentation 20, Agent ergonomics 43, Security & auth 49, Payments & pricing 30, Task success pending, Maintenance & community 3, Transparency & trust 51 · total over the 7 assessed categories
- Why: Reliability, Read with the hosted lines, on the MCP server. · Schema & documentation, Graded on the MCP server at prod.asmiai.com/mcp/claude/ (also /mcp/openai/ and /mcp/meta/). · Agent ergonomics, Graded on the MCP server. · Security & auth, Graded on the MCP server. · Payments & pricing, Read with the hosted rule. · Maintenance & community, Read for a closed service with a hosted MCP server. · Transparency & trust, Closed service under terms and a privacy policy that each say Effective October 2026, with no day, naming Humint Labs, Inc.
- Sources: 10, open questions: 3, both in the full twin
- Capabilities: assistant.messaging, assistant.memory, voice.calls, email.send, messaging.whatsapp, messaging.inbound, web.search, tasks.create, memory.user
- JSON: https://www.anchorterminal.com/api/v1/tools/asmi.json
- Verify (for the vendor): the badge `https://www.anchorterminal.com/badges/asmi.svg` or a link to https://www.anchorterminal.com/tools/asmi from a page on asmiai.com or one of its subdomains, then `POST https://www.anchorterminal.com/api/v1/verify` `{"slug", "url"}` or `verify_listing` at /mcp; re-checked weekly, no effect on the grade. Snippets in the full twin.

## Before you call it

1. Use the endpoint for your own client (/mcp/claude/, /mcp/openai/ or /mcp/meta/). A token from another client gets 403.
2. Call discover_search before execute_start, then poll execute_status with the task id instead of starting the task again.
3. Answer Asmi's questions with execute_advance. Calls already placed can't be recalled by execute_cancel.
4. Expect outreach_not_permitted if the owner unticked contact people. Ask them to reconnect rather than retrying.
5. Check billing_status before starting a task, since each outreach task uses one of the month's quota.

## Connect

```bash
claude mcp add --transport http asmi https://prod.asmiai.com/mcp/claude/
```

## Similar tools

| Tool | Grade | Score | Shared capabilities | Slim |
| --- | --- | --- | --- | --- |
| Boski | F | 26.2 | assistant.messaging, assistant.memory, messaging.whatsapp, messaging.inbound, voice.calls, tasks.create, memory.user, email.send | https://www.anchorterminal.com/tools/boski.min.md |
| Town | D | 48.7 | assistant.messaging, assistant.memory, messaging.whatsapp, messaging.inbound, voice.calls, memory.user, tasks.create | https://www.anchorterminal.com/tools/town.min.md |
| folk | F | 19.8 | assistant.messaging, assistant.memory, messaging.whatsapp, messaging.inbound, voice.calls, memory.user, tasks.create | https://www.anchorterminal.com/tools/folk-nozomio.min.md |
| szn | F | 7.3 | assistant.messaging, assistant.memory, voice.calls, email.send, messaging.inbound, memory.user, tasks.create | https://www.anchorterminal.com/tools/szn.min.md |
| Del | F | 20.6 | assistant.messaging, assistant.memory, messaging.whatsapp, messaging.inbound, tasks.create, memory.user | https://www.anchorterminal.com/tools/del.min.md |


## Panel reviews (0, desk reviews from public material, no calls made)


Disclosure: Asmi competes with LocalGhost, which Anchor Terminal's founder builds. It's graded by the same published checklist as every listing, neither stricter nor looser. Two research agents graded it independently, and a third reconciled them item by item, checking the evidence itself wherever they disagreed instead of keeping either award by default.