Zo Computer

# Fix list: Zo Computer

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/zo-computer, the October 2026 research run, assessed 12 October 2026. Grade E, 43.1 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Zo Computer: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Reliability, 15 out of 100, up to 17 more on the total

Why it scored 15: Read with the hosted lines. No status page is linked from the home page, docs index, security page or help centre home (0). With no incident history to read, 5. No rate limit is published with numbers. The changelog of 7 September 2026 says the owner can set a concurrency limit for `/zo/ask` in Advanced settings, with no default figure given (0). No 429, Retry-After, backoff or idempotency guidance for the API was found (0). No SLA, and the terms supply the service as is (0). The API is at version 1.0.0 in its OpenAPI file and is not labelled beta (10). Total 15.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 2. Security & auth, 37 out of 100, up to 11 more on the total

Why it scored 37: One bearer token created in Settings, which both guides say grants full access to the workspace, and with it the shell, every connected account and every browser session. No scopes, expiry or rotation are documented (10 of 30). Persona permissions have four presets and 19 scopes and remove tools from a conversation, and a caller can pass `persona_id`, but the token is not bound to a persona. No approval step before sending, buying or deleting was found, and the browser guide says Zo can buy and send on signed-in sites (8 of 20). The assistant reads mail, web pages and channel messages. The security page says the design is not without risks and asks researchers to report prompt injection findings, with no mitigation or guidance documented (4 of 15). Each automation run and conversation is saved with the tools it used, with no audit log of API token use documented (7 of 15). A valid security.txt and a disclosure policy with a safe harbour and a 30-day window. No bounty terms, SOC 2 or ISO 27001 report, or advisory list was found (8 of 20). Total 37.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 3. Agent ergonomics, 34 out of 100, up to 10.7 more on the total

Why it scored 34: The HTTP API is four operations, and `output_format` takes a JSON Schema so a caller can size the answer. The MCP guide says all Zo tools are exposed, and llms.txt lists 110 tool reference pages, which is well over 30. Persona permissions can remove tools, though whether the MCP endpoint honours a persona was not established (10 of 25). Structured output and streaming control output, with no pagination or filtering on the list operations (8 of 20). Errors are a validation schema and a message string with no documented codes or recovery advice (6 of 20). No idempotency key or retry guidance for `/zo/ask`, MCP annotations were not read, and the changelog for the week of 17 August 2026 records a fix for a tool call that could run twice (2 of 20). Only `input` is required and defaults are sensible. An agent-sdk repository exists under the vendor's GitHub organisation but is not mentioned in the docs we read (8 of 15). Total 34.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 4. Payments & pricing, 40 out of 100, up to 7.5 more on the total

Why it scored 40: No x402, MPP or L402 (0). Plan prices, included credit and per-request media prices are public without a login, and the billing guide says chat follows listed token rates (20). The Free Trial runs 14 days and the home page says no credit card is required (20). A person signs up in a browser behind a Turnstile check and creates the token in Settings, so an agent cannot get access by itself (0). Total 40.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 5. Schema & documentation, 73 out of 100, up to 4.4 more on the total

Why it scored 73: Graded on the HTTP API, with the MCP endpoint's tool definitions unread because listing them needs a token. A public OpenAPI 3.1 file covers all four operations (25). llms.txt and a Markdown copy of every guide and reference page (10). Field and tool descriptions state purpose in one line and seldom say when not to use something (10 of 20). The ask request is typed with one required field and an enum for `memory_mode`, while the work itself is a free-text prompt (10 of 15). The guide has curl and Python examples for continuing a conversation, structured output and streaming. Documented errors are a 422 validation schema, an error response schema and an SSE `Error` event, with no list of codes (8 of 15). A dated weekly changelog is public and the spec has a version, but the paths carry none and the guide omits `memory_mode`, which the spec defines (10 of 15). Total 73.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 6. Transparency & trust, 60 out of 100, up to 3.5 more on the total

Made of editorial 46, provenance 73.

Why it scored 60: Closed source with public terms. The terms are 18 short clauses last updated on 8 August 2025, cite an Acceptable Use Policy we did not find, name "Substrate Lab's trademarks" in clause 3, and say credits are bought in advance while the product is sold by subscription (12 of 30). The privacy policy of 20 May 2026 lists only email address, trackers, usage data and SMS data, gives no retention period, and does not describe workspace files, memory or connected-account data. The security page fills part of that gap, with stated retention for Postmark (45 days) and Telnyx (10 days), and the help centre says deleting the account erases the computer. No DPA was found (14 of 30). The changelog announces retired and deprecated models and says retired models forward to a successor, with no deprecation policy or notice period for the API (6 of 20). The security page names each provider and marks which can see user content. Data locations are given only for Fireworks, as US-based (14 of 20). Total 46.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: zo.computer, no registry record we could read (0 of 15)
- Terms of service: read, states 5 of the 7 things a reader expects (8.3 of 10)
- Privacy policy: read, states 7 of the 8 things a reader expects (9.3 of 10)
- Status page: not found (0 of 10)

## 7. Maintenance & community, 68 out of 100, up to 2.8 more on the total

Why it scored 68: The newest dated changelog entry is the week of 14 September 2026, 28 days before this check, followed by a September summary. No entry for the three weeks since was on the page (30). Nine weekly entries fall in the last 90 days (20). A public changelog, a Discord community and a help centre with a route to a person. We did not test whether they answer (9 of 15). No official SDK is documented for the API and the official MCP registry returned no server for zocomputer. The desktop app's last GitHub release is v1.5.13 of 13 August 2026 (4 of 15). Package health does not apply to a closed service, and half marks rest on the changelog's steady run of fixes (5 of 10). Total 68.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- Kind chosen: `http-api`. Zo is a hosted assistant, which the brief would call a platform, but it has a documented public HTTP API with an OpenAPI file and an MCP endpoint, so it was graded as a hosted service with an API. The MCP endpoint is not the main surface for a person, who uses chat and messaging
- `assistant.self-hosted` was not claimed. The vendor describes a personal server and self-hosting, but the server runs on Modal or Daytona under the vendor's account
- Whether the assistant asks before sending mail, posting or buying. No approval step was found in the docs read, and the product was not run
- Which OAuth scopes each connected app requests and where the tokens are stored. The security page says Pipedream has no data access while the tool reference says OneDrive and Dropbox work via Pipedream
- What memory stores, where, and whether one memory can be deleted without a factory reset
- Whether an access token can be revoked individually, and whether the MCP endpoint honours persona permissions
- The terms cite an Acceptable Use Policy that was not found, and clause 3 names Substrate Lab
- unchecked: the MCP tool definitions, count and annotations. Listing them needs a token and nothing was sent to api.zo.computer
- unchecked: per-token chat prices on the models pages, and the default `/zo/ask` concurrency limit
- unchecked: the remaining guide pages (tools, skills, bring your own key, devices, payments, hosting) and the 110 tool reference pages other than Use Gmail and List available scopes, which were left unread to limit requests. We made 21 requests to www.zo.computer, above the brief's guide of about fifteen
- unchecked: the zocomputer/agent-sdk and zocomputer/skills repositories, GitHub security advisories for the organisation, and the domain's registration date
- unchecked: whether support, Discord or the help centre answer, and whether a status page exists at an address no page links
- The changelog page shows no entry after the week of 14 September 2026. `lastRelease` is that week's Monday

## Weaknesses

- The access token for the API and MCP endpoint grants full access to the workspace, with no scoped or read-only token documented
- No approval step before sending mail, posting or buying was found. The browser guide says Zo can buy and send on signed-in sites
- No status page, incident history, rate limit figures or SLA were found on the pages read
- The privacy policy lists only email address, trackers, usage data and SMS data, and does not describe workspace files, memory or connected accounts
- Memory storage and deletion are not documented beyond a design aim and whole-account deletion or factory reset

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Send `POST https://api.zo.computer/zo/ask` with `Authorization: Bearer zo_sk_...` and an `input` string. The owner creates the token under Settings > Advanced
- Pass `persona_id` for a persona with a narrow tool set. The token itself carries every permission the owner has
- Set `memory_mode` to `off` to keep workspace memory out of a request. The field is in the OpenAPI file and absent from the API guide
- Do not retry `/zo/ask` blindly after a timeout. No idempotency key is documented and the request may already have sent a message
- Append `.md` to any guide or reference URL for Markdown

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.