Head to head · Cad generate · October 2026 research run
Adam vs Zoo
Adam and Zoo score within a point of each other on agent readiness, 59.4 (C) and 58.7 (C). Zoo leads on reliability, payments & pricing and maintenance & community. Both do cad generate.
Best CAD, simulation and fabrication APIs for AI agents · All 15 cad comparisons
Which one, for what
Adam C
Best for An agent that drives mechanical CAD work in Onshape, Fusion or SOLIDWORKS through a hosted engineering agent, or needs STEP and 3MF exports of Adam-stored models.
Ahead on
- Schema & documentation, 81 against 65
- Agent ergonomics, 82 against 44
- Security & auth, 74 against 67
Also in its favour
- A hosted endpoint, with nothing to install
Watch for
The MCP server and REST API are labelled beta, with breaking changes allowed
Zoo C
Best for Agents that need to produce or iterate editable mechanical geometry as code and export STEP, with a hosted engine and no desktop.
Ahead on
- Reliability, 55 against 35
- Payments & pricing, 35 against 15
- Maintenance & community, 90 against 55
Also in its favour
- Runs on your own machine
Watch for
No rate limits, 429 handling or idempotency keys documented for the REST API or the websocket
Why the edge
Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.
Agent ergonomics Adam +38
Adam 82
The docs name 15 adam_ tools plus displayFile and loadSkill, so 11 to 30 (15), and task tools are discovered per task with adam_list_task_tools rather than listed up front, which earns 8 back (23 of 25). Cursor pagination with limit 1 to 50, status and project filters on tasks, and message parts that admins can trim, though results of tool calls may be truncated without a size control (18 of 20). One error envelope with type, message and requestId, 402 and 403 that name the plan or scope at fault, and a recovery table for each direct-tool failure (18 of 20). Idempotency-Key required on every write that starts work, a 24-hour replay, cancel documented as safe to repeat, and an unknown status that tells the agent to inspect before retrying; deletes need confirmation, but readOnlyHint and destructiveHint annotations were not confirmed since tools/list was not read (16 of 20). Few required parameters, but no official SDK in any language (7 of 15). 23 + 18 + 18 + 16 + 7 = 82.
Zoo 44
zoo-mcp lists 42 tools in one process with no toolset switch, read-only subset or dynamic loading, and their docstrings total about 30,000 characters (5 of 25). List endpoints take limit, page_token and sort_by (31 paginated operations), with no field selection or filtering (15 of 20). Errors are a generic object with a message and request_id, no error codes, and MCP tools return a string on failure (8 of 20). No idempotency keys, no readOnlyHint or destructiveHint in the tool source, and the websocket needs a ping every 5 seconds to stay open (4 of 20). Few required parameters, official SDKs in Python, TypeScript, Rust and Go, but one modeling session per MCP process and a websocket protocol an agent has to drive by hand (12 of 15). 5 + 15 + 8 + 4 + 12 = 44.
Maintenance & community Zoo +35
Adam 55
Read for a closed service. The changelog's newest entry is 10 September 2026, 54 days before this check (20). Entries dated 7, 10, 11 and 25 August and 4 to 10 September 2026, more than three in 90 days (20). A public changelog, hello@adam.new, a Discord and the CADAM repository with 26 open issues, which were not read for replies (10 of 15). The manifest follows the registry schema but new.adam/adam is not in the official MCP registry, and there is no official SDK (0 of 15). CADAM was pushed on 15 October 2026 and the service itself has no package to judge (5 of 10). 20 + 20 + 10 + 0 + 5 = 55.
Zoo 90
Read for a hosted service with open-source clients. zoo-mcp v0.33.0 on 30 October 2026, the CLI v0.2.209 on 28 October and kittycad 1.8.0 on 15 October (30). Nine zoo-mcp tags from 24 to 30 October alone, with CLI and SDK releases through October (20). Commits land daily across the KittyCAD organisation, with a Discord and a community forum; issue response times were not read because GitHub's issue pages refused us (15 of 25). io.github.KittyCAD/zoo-mcp is in the official MCP registry under a GitHub-verified namespace, though at version 0.10.0 of 27 January 2026 (15). pytest across Python 3.11 to 3.14, ruff, ty, a security workflow and Dependabot, with the scheduled run on main passing on 3 November (10). 30 + 20 + 15 + 15 + 10 = 90.
Reliability Zoo +20
Adam 35
Read with the hosted lines. status.adam.new exists and is linked from the docs, but it loads its components and history by script and nothing readable was served, so the page counts for half (10 of 20) and the incident record is unread (5). Rate limits are documented as two refilling buckets reported in RateLimit, RateLimit-Policy and X-RateLimit headers, with no numeric quota published (5 of 15). 429 carries Retry-After, and Idempotency-Key is required on task creation, messages and tool calls and replayed for 24 hours (15). No SLA found (0). The MCP server and the REST API are labelled beta, with breaking changes allowed (0 of 10). 10 + 5 + 5 + 15 + 0 + 0 = 35.
Zoo 55
Read with the hosted lines on api.zoo.dev. A Better Stack status page at status.zoo.dev lists public and internal components with uptime figures (20). Its incident pages show no incidents for September, October and November 2026; the June to August archive didn't render for our reader, so about three weeks of the 90-day window are unread (25 of 30). No rate limit with a number was found in the OpenAPI document, the API docs, the FAQ or the MCP source (0). No Retry-After or backoff guidance and no idempotency keys; the document says only that cancellation is idempotent (3 of 15). No SLA on any plan, and the terms disclaim uninterrupted service (0). The REST API and Zookeeper are not labelled beta, while POST /ml/kcl/completions carries the beta tag and zoo-mcp declares the Beta classifier (7 of 10). 20 + 25 + 0 + 3 + 0 + 7 = 55.
Payments & pricing Zoo +20
Adam 15
Read with the hosted rule. No x402, MPP or L402 (0). The pricing page's prices are drawn by script and were not readable; the Markdown version and the billing docs name Free, Pro, Max, Team and Enterprise with no figure, so pricing is scored as unread (0 of 20, named in openQuestions). A Free plan with a daily usage pool and Lite models exists, and trials charge a saved card only when confirmed; whether Free needs a card is not stated (15 of 20). A person signs in with OAuth in a browser and picks a workspace (0). 0 + 0 + 15 + 0 = 15.
Zoo 35
No x402, MPP or L402 (0). The per-second rate ($0.0083) and plan allowances are public in the pricing JSON the site loads and on the blog, with the FAQ stating rounding and free GETs (20). The Free plan carries $10 of usage a month with no plan fee; whether sign-up asks for a card was not established, and pay-as-you-go needs a payment method (15 of 20). A person signs in with SSO in a browser and mints the token; service accounts come only after an organisation exists (0). 0 + 20 + 15 + 0 = 35.
Schema & documentation Adam +16
Adam 81
Public OpenAPI 3.1 for the REST API (adam.new/openapi.json and api.adam.new/openapi.json), a site spec for OAuth and MCP endpoints, and an MCP manifest. MCP tool schemas were not read, since tools/list needs a sign-in, though adam_list_task_tools is documented to return input schemas (20 of 25). llms.txt, a docs index at docs.adam.new/llms.txt, Markdown for every page on request through the Accept header, and a starter SKILL.md (10). The direct-tools page gives each MCP tool a purpose and its scopes, and OpenAPI descriptions state conditional requirements such as cron or runAt, but no tool definition says when not to use it (12 of 20). Enums for task status, limit 1 to 50, required fields and multipart file fields are typed; the component schemas past the first 100,000 characters were not read (11 of 15). Every operation documents 400 to 500 responses with one error envelope, and the API introduction carries curl and JSON examples (13 of 15). Versioned /v1 paths, Deprecation and Sunset headers with 90 days' notice, and a dated changelog, though its newest entry is 10 September 2026 (15). 20 + 10 + 12 + 11 + 13 + 15 = 81.
Zoo 65
OpenAPI 3.0.3 served at the root of api.zoo.dev, 174 paths and 226 operations each with a summary, and MCP tools whose inputs are typed Python signatures that the SDK turns into JSON Schema (25). No llms.txt (404) and no Markdown twin of the docs (0). The MCP docstrings average about 725 characters and say what each tool is for and when to use another, with snapshot at 3,500 characters; the OpenAPI summaries are one line and the websocket page lists one client message type (14 of 20). 583 schemas with 976 enum lists and required fields, and the file conversion path takes typed format parameters (12 of 15). Each docs endpoint page carries a request example, but errors are one generic object with request_id and message, and the websocket's server messages are not enumerated in the docs (9 of 15). The document reports version 0.1.0, there is no API changelog page, and changes are announced per repository release or on the blog, as the per-second billing change of 30 September 2025 was (5 of 15). 25 + 0 + 14 + 12 + 9 + 5 = 65.
Security & auth Adam +7
Adam 74
OAuth 2.1 authorisation code with PKCE S256, automatic client registration, RFC 8707 resource indicators, refresh tokens, a revocation endpoint and 11 scopes, or a personal key minted for one workspace with chosen scopes and shown once; bearer header only, no query-string option (30). Read and change approved separately per area, tools:execute kept apart and blockable by an admin under Allow full harness execution, deletes behind confirmation, and access revoked at once from Connected apps. tools:execute is broad, covering commands, connected accounts, paired devices, stored credentials and billable tools (18 of 20). Task messages return the agent's output and tool results, which can include web search and connected-account content, and no prompt-injection guidance was found; per-user sandboxes isolate execution but do not address it (4 of 15). Connected apps are listed with revocation, every error carries a requestId and the billing summary reports usage, but no audit log for the operator was found (6 of 15). SOC 2 Type I approved with Type II in progress, a bug bounty of $10 to $100 per finding with a 90-day window and safe harbour, patch SLAs by CVSS severity, 72-hour incident notification, and no security.txt (16 of 20). 30 + 18 + 4 + 6 + 16 = 74.
Zoo 67
Bearer API tokens minted in the account, several per user, listed and deleted at /user/api-tokens, organisation service accounts on Team and Enterprise, and OAuth 2.0 apps with PKCE and a device grant; no scopes were found on tokens and no rotation guidance (22 of 30). No read-only token or MCP subset; the Team plan's admin console and auditing give oversight after the fact, and the MCP client's own file permissions govern writes to main.kcl (8 of 20). Zookeeper carries a web_search tool and the MCP server returns KCL samples and docs, with no prompt-injection guidance found (5 of 15). Organisation admins get an API activity log and /user/api-calls lists each call with its cost (12 of 15). security.txt with a GitHub advisory contact and a policy committing to a 72-hour reply and severity-based rewards, SOC 2 Type II on the trust centre, and two Trail of Bits reviews published in 2024 and 2026 (20 of 20). 22 + 8 + 5 + 12 + 20 = 67.
Score by category
| Category | Weight this run | Adam | Zoo | Edge |
|---|---|---|---|---|
| Reliability | 16%20 | 35 | 55 | Zoo +20 |
| Performance | 10%pending | pending | pending | not scored in this run |
| Schema & documentation | 13%16.2 | 81 | 65 | Adam +16 |
| Agent ergonomics | 13%16.2 | 82 | 44 | Adam +38 |
| Security & auth | 14%17.5 | 74 | 67 | Adam +7 |
| Payments & pricing | 10%12.5 | 15 | 35 | Zoo +20 |
| Task success | 10%pending | pending | pending | not scored in this run |
| Maintenance & community | 7%8.8 | 55 | 90 | Zoo +35 |
| Transparency & trust | 7%8.8 | 72 | 69 | Adam +3 |
| Negative events | ≤15 | 0 | 0 | |
| Total | C 59.4/100 | C 58.7/100 |
Facts side by side
| Fact | Adam | Zoo |
|---|---|---|
| Kind | MCP server | HTTP API |
| Vendor | Adam AI Labs, Inc. | Zoo Corporation |
| Hosted endpoint | https://adam.new/mcp | no (local only) |
| Transports | Streamable HTTP | stdio, Streamable HTTP |
| Auth | OAuth or key | OAuth or key |
| Pricing | Freemium | Freemium |
| x402 | no | no |
| Licence | Closed service under Adam's Terms of Service (Adam AI Labs, Inc., last updated 5 June 2026). The CADAM text-to-CAD web application is open source under GPL-3.0. | none |
| Tools exposed | none | 42 |
| Read-only variant documented | no | yes |
| llms.txt | yes | no |
| MCP registry | not listed | io.github.KittyCAD/zoo-mcp |
| Last release | 2026-09-10 | 2026-10-30 |
| Terms last updated | couldn't be read | 2026-04-15 |
| Privacy policy last updated | couldn't be read | 2026-09-09 |
| Customer content may train models | couldn't be read | yes, with an opt-out |
| Terms restrict automated access | couldn't be read | yes |
| Terms restrict benchmarking | couldn't be read | yes |
| Terms or service can change without notice | couldn't be read | not found in the text |
| Arbitration or class-action waiver | couldn't be read | not found in the text |
| Popularity | 5.4k stars | 11 stars |
Verdicts
Adam
An agent reaches Adam's CAD tasks over OAuth 2.1 with per-area scopes, idempotency keys and a public OpenAPI spec, and can export STEP, STL, IGES and 3MF from Adam-stored models. The MCP server and API are beta, numeric rate limits and plan prices are not published, and native CAD edits need a connected desktop application.
Zoo
A public OpenAPI document, SDKs in four languages, a 42-tool MCP server and a $10 monthly allowance let an agent generate KCL geometry and export STEP without a contract. No rate limits, retry guidance or scoped tokens are documented, the MCP server has no read-only subset, and the Free plan's data trains Zoo's models with no opt-out.
Before you call either
Adam
- Call adam_get_context first and check the scopes; client-driven tool calls need
tasks:read,tasks:writeandtools:execute, which an admin may have blocked - Send a fresh UUID as Idempotency-Key on every task, message and tool call, reuse it on a retry, and on status unknown inspect the task before retrying with a new key
- Poll adam_get_tool_call for queued or running calls; creating an external task starts no inference, but executing tools spends the workspace's usage
- Treat 402 as a plan or usage limit for the person to resolve, and on 429 wait for Retry-After
- Expect only Adam-stored models to export without a desktop; ask the person to connect Onshape, Fusion or SOLIDWORKS before native edits
Zoo
- Send
{"type":"ping"}every 5 seconds on the Zookeeper websocket and expect only KCL back; run the KCL through evaluate_kcl, the CLI or zoo-kcl to get STEP or STL - Budget usage in seconds, since every billable call rounds up to the next second at $0.0083, and the monthly allowance resets at the account anniversary without rollover
- Use one modeling session per zoo-mcp process, start it with start_modeling_session and stop it when done; get_modeling_sessions recovers the id after a reconnect
- Put the token in ZOO_API_TOKEN, never in a URL; there are no read-only tokens, so use a service account on a Team plan when the owner wants separate credentials
- Paginate list endpoints with
limit,page_tokenandsort_by; nothing documents a rate limit, so back off on any 429 or 5xx yourself
Questions
Which is better for AI agents, Adam or Zoo?
Adam and Zoo score within a point of each other on agent readiness, 59.4 (C) and 58.7 (C). Zoo leads on reliability, payments & pricing and maintenance & community.
Do Adam and Zoo need an API key?
Both take an API key or an OAuth sign-in.
Can an agent call Adam and Zoo without installing anything?
Adam has a hosted endpoint at https://adam.new/mcp. Zoo runs on your own machine, with no hosted endpoint listed.
Other comparisons with Adam or Zoo
Machine-readable
- This page as Markdown
/compare/adam-cad-vs-zoo.md· slim.min.md· JSON.json(or sendAccept: text/markdown) - Each listing in full
/api/v1/tools/adam-cad.json·/api/v1/tools/zoo.json - From a terminal
anchor compare adam-cad zoo(the CLI) - Over MCP
compare_tools {"a": "adam-cad", "b": "zoo"}at/mcp, no key