0m · surface
Every model.
One bill you
actually control.
StrataBlock is one endpoint for the leading AI models, with hard budgets that stop overspend the moment it happens, usage you can attribute to the request, and clear per-token prices.
scroll to descend
layer 01 · 8m · the catalog bed
Every lab that matters, one key
Frontier models next to the best open weights, served through one OpenAI-compatible API your tools already speak. Your admin curates the catalog; new models land as the labs ship them.
one band per lab, drawn to scale of ambition
layer 02 · 16m · the interface seam
Change one line, keep your stack
Point your OpenAI SDK at StrataBlock and everything keeps working: streaming, tool calls, structured output, vision.
curl https://your-strata-host/v1/chat/completions \
-H "Authorization: Bearer $STRATA_API_KEY" \
-d '{"model": "claude-sonnet", "messages": [{"role": "user", "content": "Hello"}]}'from openai import OpenAI
client = OpenAI(
base_url="https://your-strata-host/v1",
api_key=STRATA_API_KEY,
)
reply = client.chat.completions.create(
model="claude-sonnet",
messages=[{"role": "user", "content": "Hello"}],
)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://your-strata-host/v1",
apiKey: process.env.STRATA_API_KEY,
});
const reply = await client.chat.completions.create({
model: "claude-sonnet",
messages: [{ role: "user", content: "Hello" }],
});layer 03 · 24m · the control stratum
Built for teams that watch the bill
find 01
Budgets that stop spend, not report it
Set a hard cap for the org and for each person. At the cap, requests stop with a clear error in real time, not a surprise on next month's invoice.
find 02
Attribution down to the request
Scoped keys per app and environment, freeform tags on every call, and dashboards you can slice by model, person, key, tag, and day. Export it all to CSV.
find 03
Spend you can actually read
Live usage, per-token prices, and a dashboard that matches the bill. What you see is what comes off your balance.
layer 04 · 32m · embedded toolbeds
More than an API
tool 01
StrataChat
A full chat workspace with projects, memory, and built-in web search.
tool 02
StrataBench
Benchmark models side by side on your own prompts.
tool 03
StrataSwarm
Multi-agent deep research with a replayable execution trace.
solid ground · 40m · plain numbers
Plain pricing, on the page
A flat monthly fee for the platform. Usage comes from prepaid credits at clear per-token prices. No invoices lurking behind a sales call.
Team
- •Up to 10 people
- •50 API keys
- •120 requests/min
- •Prepaid credits for usage
Team Plus
- •Up to 25 people
- •150 API keys
- •200 requests/min
- •Prepaid credits for usage
Team Pro
- •Up to 50 people
- •400 API keys
- •300 requests/min
- •Prepaid credits for usage
Enterprise
Unlimited members and keys, run managed or self-hosted in your own cloud.
fossil layer · 48m · field reports
What teams say
tag no. 001
"We rolled StrataBlock out to the whole studio in an afternoon. Every engineer got every model, and finance stopped asking me what the AI line item was."
tag no. 002
"An agent went into a retry loop on a Friday night. The budget cap caught it at the limit and just stopped it. That one evening paid for the year."
tag no. 003
"One key per app, tags on every call. For the first time our AI spend has names on it, and the CSV drops straight into our month-end reporting."
copper vein · runs on your land
StrataBlock Enterprise
The same gateway, self-hosted in your own cloud account. Your data never leaves your perimeter.
field notes · 56m
Questions, answered
What does OpenAI-compatible actually mean here?
StrataBlock implements the OpenAI Chat Completions and Responses APIs, including streaming, tool calls, structured outputs, and multi-modal content where the model supports it. Any client library that targets api.openai.com works against your StrataBlock base URL with no code change beyond the URL and the key.
How do API keys and tags work together?
Mint any number of scoped keys, typically one per app, service, or environment. Attach a freeform tag string to any request, such as project=growth,env=staging. The usage dashboard then groups and filters spend by any combination of key, model, and tag, so attribution down to a single experiment is one click away.
How are rate limits and budgets enforced?
Per-key rate limits sit in front of every model, so a runaway loop in one app cannot starve the others. Budgets are checked server-side on every call: at the cap, new requests stop immediately with a clear error. You see the breach in the same request, not in tomorrow's billing email.
Where does my data live? What gets stored?
Your traffic stays inside AWS data centers from the moment it reaches the gateway. We never store the contents of your requests or responses. The only data we keep is operational metadata for monitoring and attribution: model, token counts, latency, status, and your tags. No prompts. No completions.
How does billing work?
A flat monthly platform fee plus prepaid credits for usage. Every request is priced per token at rates you can see in the dashboard, and usage rolls up by model, person, and day with CSV export. The number in the dashboard is the number you pay.
How fast can I start?
Sign up, create a key, point your SDK at the endpoint. Most people make their first call in minutes; there is no waitlist and nothing to install.
resurfacing · back at the surface
Give your team every model today
Set a budget, hand out keys, and stop guessing at the bill.
Start now