scanned Oct 18, 2026
typesafe.ai
TypeSafe AI builds machine-native intelligence infrastructure for automation
Tier 3 · Agent-Accessible
Scored by asking 14 questions a buyer of a ai-ml product asks, then grading this site’s own pages: answered, hedged (partial or vague), or silent (no page answers it). How scoring works
14 buyer questions · graded against the site’s own pages
Goes silent · integration · medium importance
Does the console support SAML-based single sign-on and automated user provisioning through SCIM 2.0?
Goes silent · security · medium importance
Are console and API audit logs immutable, and can they be streamed in real time to an external SIEM in JSON format?
This report is public. Own typesafe.ai? Claiming is free: crawl every page, re-audit as you fix, and track your score over time.
Sign in to claim49 points sit between typesafe.ai and 100: 10 open questions and 8 missing protocol checks, ordered by estimated payoff.
Point estimates are per fix under scoring v2. They are not additive to a promised total.
Does the console support SAML-based single sign-on and automated user provisioning through SCIM 2.0?
No page on the site addresses this.
Add a dedicated help center article or documentation page that lists console authentication and provisioning capabilities, explicitly stating SAML SSO and SCIM 2.0 support or roadmap.
confidence high · grounding world-knowledge · weight 0.00 · Absent
Are console and API audit logs immutable, and can they be streamed in real time to an external SIEM in JSON format?
No page on the site addresses this.
Create a dedicated security or compliance page that documents audit log capabilities, including immutability guarantees, retention policies, and SIEM integration options such as real-time JSON streaming endpoints or webhook configurations.
confidence high · grounding world-knowledge · weight 0.00 · Absent
What is the guaranteed response time for critical production incidents under the Enterprise support plan?
No page on the site addresses this.
Create a dedicated Support or SLA page that lists response time commitments by plan tier, including the guaranteed response time for critical production incidents under the Enterprise plan.
Where we looked: /enterprise
confidence high · grounding synthesized · weight 0.00 · Absent
How long do you guarantee backward compatibility for API versions after releasing a breaking change?
No page on the site addresses this.
Publish a dedicated API versioning and deprecation policy that specifies how long previous API versions remain available after a breaking change.
Where we looked: /sdk/javascript/changelog.md, /sdk/python/changelog.md, /blog/jev-1-14
confidence high · grounding synthesized · weight 0.00 · Absent
A buyer gets something, then has to guess the rest. Expand any row for the evidence and the fix.
Can the inference stack be deployed in our own VPC or on-premises, or is it strictly available as a multi-tenant SaaS?
No page on the site addresses this.
Expand the FAQ accordion for 'Can we run Jev in our own VPC or on-premises?' on the enterprise page so the crawler captures the answer text.
Where we looked: /enterprise
confidence high · grounding synthesized · weight 0.00 · Collapsed
Do you offer a pay-as-you-go developer tier, or is access only available through annual enterprise contracts?
Typesafe AI's documentation examples reference a 'cloud edition's self-serve Team plan' and an 'Enterprise contract,' and the contact page lists 'Individual Builders' alongside 'Enterprise' and 'Partnerships.' However, none of the retrieved chunks specify whether the Team plan is pay-as-you-go, or whether enterprise contracts are exclusively annual.
Add a public pricing page that clearly outlines available tiers (e.g., Individual/Team/Enterprise), billing models (pay-as-you-go vs. annual contracts), and self-serve signup options.
Where we looked: /cookbooks/basic_memory.md, /cookbooks/basic_memory, /enterprise, /contact
confidence medium · grounding synthesized · weight 0.65 · Page missing
If we exceed our included monthly inference quota, are requests throttled or billed as overage at a per-token rate?
The retrieved chunks do not mention monthly inference quotas or overage billing. The site documents per-organization rate limits measured in tokens per second and requests per second; exceeding these returns a 429 Too Many Requests error. Higher limits are available on custom and enterprise plans.
Add a pricing or plan limits page that explains monthly inference quotas and whether exceeding them triggers throttling or per-token overage billing.
Where we looked: /api.md, /api, /models.md
confidence medium · grounding synthesized · weight 0.65 · Page missing
When a System One structured output fails schema validation, does the platform automatically retry generation, and what is the maximum retry count?
The retrieved chunks do not indicate that the platform automatically retries generation specifically when a System One structured output fails schema validation. The documented automatic retries apply to HTTP errors such as 429, 500, 502, 503, 504, and 529, plus connection failures and timeouts. The default maximum retry count is 2 retries after the initial attempt, and it is configurable via RetryPolicy—examples show settings such as max_retries=3.
Add explicit documentation on whether schema validation failures for System One structured outputs trigger automatic retries, and if so, how they interact with RetryPolicy.
Where we looked: /sdk/python/usage/basic, /sdk/python/api/retries.md, /sdk/python/api/retries, /api, /api.md
confidence medium · grounding synthesized · weight 0.65 · Vague
What are the exact requests-per-minute rate limits for the standard tier, and do they differ for coding agents versus System One models?
The chunks do not provide exact requests-per-minute rate limits for a standard tier. Chunk [11] states that rate limits are set per organization and measured in tokens per second and requests per second (not per minute); current limits are available in the developer console, and higher limits are available on custom and enterprise plans. None of the chunks indicate that rate limits differ for coding agents versus System One models; coding agents are mentioned only in integration and batching contexts, not in relation to rate limits.
Add exact rate limit numbers (requests per minute/second) for each plan tier to the Models or Pricing page, and clarify whether coding agents and System One models share the same limits or have separate quotas.
Where we looked: /models.md
confidence high · grounding synthesized · weight 0.70 · Vague
How long do you retain prompt and completion logs by default, and can we configure a zero-retention policy for sensitive data?
Zero Data Retention (ZDR) is available by request to enterprise customers; when enabled, only operational telemetry is retained. However, the site does not specify a default retention duration for prompt and completion logs.
Add the default retention period (e.g., 30 days, 90 days) for prompt and completion logs to the data retention help article or ZDR page.
Where we looked: /articles/9520184411-do-you-offer-zdr, /articles/5974925716-how-is-submitted-data-used-or-retained
confidence high · grounding synthesized · weight 0.70 · Vague
4 of 14. Answered with enough specificity for a buyer to act on. This is the bar the queue above is chasing.
Which programming languages have officially supported SDKs or client libraries for your API?
TypeSafe officially supports Python and JavaScript SDKs.
Source: /sdk/javascript, /sdk/python
confidence high · grounding stated · weight 1.00 · Answered
Is your platform SOC 2 Type II certified, and do you support HIPAA or GDPR data processing agreements?
TypeSafe maintains SOC 2 Type II certification. It publishes a Data Processing Addendum (DPA) that includes EU Standard Contractual Clauses and the UK Addendum to support GDPR data processing. However, TypeSafe does not currently support HIPAA/BAA; it is working toward making its Jev product available through platforms like Amazon Bedrock, Azure Foundry, and Google Vertex AI for teams with BAA/HIPAA requirements, but there is no confirmed availability date.
Source: /articles/8595205272-can-typesafe-provide-security-documents-or-complete-our-vendor-assessment, /articles/9159694888-do-you-offer-a-dpa-baa-scc-and-specific-data-residency-options, /legal
confidence high · grounding synthesized · weight 0.95 · Answered
Is your API fully compatible with OpenAI's SDK format, or does migration require rewriting request schemas and error handling?
The chunks do not indicate that TypeSafe's API is fully compatible with OpenAI's SDK format. Instead, they document a proprietary SDK with distinct request schemas—such as the `systemOne()` method accepting `state` and `questions` objects—and proprietary error classes including `APIError`, `APIConnectionError`, and `APITimeoutError`. Chunk 1 shows the OpenAI Python client being imported and instantiated separately alongside `TypeSafeClient`, suggesting the two SDKs are used side-by-side rather than as drop-in replacements. Migration would therefore require rewriting request schemas and error handling.
Source: /cookbooks/sde_cascade.md, /llms.txt, /introduction/quickstart.md, /sdk/javascript/api/classes/TypeSafeClient.md
confidence high · grounding synthesized · weight 0.95 · Answered
What is the maximum number of concurrent primitives or total payload size allowed in a single API request?
There is no explicit limit on the number of concurrent primitives (questions) in a single API request; the constraint is that the request must fit the model's per-request context and per-question context. The total payload body size is limited to at most 32 MiB (headers do not count).
Source: /primitives.md, /api
confidence high · grounding synthesized · weight 0.95 · Answered
The other half of the score: 16 checks for the files and headers agents look for. The 8 below are installs, not judgment calls, and most are an afternoon. Expand any for the snippet and the standard it follows. They sit after the queue because none of them changes what your pages say.
Sitedex generates this file from your crawl. Grab it in Files from this audit below.
Standardllmstxt.orgCommunity spec
User-agent: * Content-Signal: search=yes, ai-input=yes, ai-train=no Allow: /
StandardCloudflare proposalVendor proposal
Sitedex generates this file from your crawl. Grab it in Files from this audit below.
StandardModel Context ProtocolCommunity spec
StandardOpenAPI SpecIndustry standard
Sitedex generates this file from your crawl. Grab it in Files from this audit below.
StandardW3C WebMCP draftW3C / WHATWG
<meta name="description" content="typesafe.ai: [outcome you deliver] for [who you help]. One sentence, 50-160 characters." />
StandardHTML Living StandardW3C / WHATWG
Sitedex generates this file from your crawl. Grab it in Files from this audit below.
StandardSchema.org + JSON-LDIndustry standard
Standardsitemaps.orgIndustry standard
Already passing 8 of 16: robots.txt, sitemap.xml, AI crawler access, Clean crawl, Markdown negotiation, Server-rendered content, Canonical URLs, HTML lang attribute.
Sitedex already serves typesafe.ai as an MCP endpoint. Ask typesafe.ai anything an AI agent might ask, and see what its index returns. (To score your own site, use the form below.)
For developers and the engineer-on-call: copy these into your tools or your site.
Files from this audit
Built from this crawl. Download or copy each, then install it at the path noted.
Built from this crawl. Install at /llms.txt so agents start here.
Organization JSON-LD, pre-filled from this crawl. Wrap in a ld+json script.
MCP server card built from this crawl. Host at /.well-known/mcp/server-card.json.
WebMCP discovery manifest built from this crawl. Host at /.well-known/webmcp.json.
MCP endpoint
https://mcp.sitedex.dev/s/typesafe-ai/mcp
The URL anyone's agent points at. Read-only; safe to share.
Claude Code
claude mcp add typesafe --transport http https://mcp.sitedex.dev/s/typesafe-ai/mcp
One command, then the agent has it.
Cursor / Continue
{
"mcpServers": {
"typesafe": {
"url": "https://mcp.sitedex.dev/s/typesafe-ai/mcp"
}
}
}Drop into mcp.json.
WebMCP: two parts
WebMCP-capable browsers run the widget at runtime. Crawlers without JS rendering need the discovery manifest to find your tool surface. Install both.
1 · Widget script
<script async src="https://sitedex.dev/widget.js"></script>
Drop in <head>. In browsers with WebMCP, this script registers your site's tools with navigator.modelContext.registerTool().
2 · Discovery manifest
{
"schema_version": "2026-04",
"name": "TypeSafe AI",
"description": "TypeSafe AI builds machine-native intelligence infrastructure for automation. WebMCP tools served by Sitedex.",
"homepage": "https://typesafe.ai",
"documentation": "https://app.sitedex.dev/sites/typesafe.ai",
"source": "navigator.modelContext",
"tools": [
{
"name": "search_site_content",
"description": "Ask about this site in plain language and get the most relevant passages from its pages, each with the source URL and a relevance score. Searches the site's published text, not live inventory or accounts."
},
{
"name": "get_site_info",
"description": "This site's profile: name, description, category, who it is for, page count, and AI Visibility Score. Use it to learn what the site is before you search it."
},
{
"name": "list_site_pages",
"description": "The pages indexed on this site, with title, URL, and path. Use it to see what the site covers, or to find the path to pass to get_site_page."
},
{
"name": "get_site_page",
"description": "The full markdown of one page on this site, by path. Use it after search_site_content points you at a page and you want all of it."
}
]
}Host alongside the script at /.well-known/webmcp.json. Crawlers that don't render JS rely on this.
Your turn
See which of these questions your site goes silent on.
Free, about 5 minutes. We crawl your site, test it against the buyer questions your category asks, and name what’s vague, contradictory, or missing, plus the files AI agents look for.