# A model built for deterministic tasks

URL: https://interfaze.ai/

Backed by

![Y Combinator](https://interfaze.ai/yc_logo.svg)

Combinator

Accurate, verifiable output in one call.

A new model architecture for OCR, audio understanding, structured data extraction and more, built for the highest accuracy, precision and consistency.

[Build now](https://interfaze.ai/dashboard/playground)Setup with agents

-   Consistency
    
    #1 on consistency and vision benchmarks
    
-   Verifiable
    
    Confidence scores, evidence, bounding boxes
    
-   Multimodal
    
    Text, images, audio, files and video
    
-   Configurable
    
    Truly controllable beyond prompting with type-safety
    
-   Multilingual
    
    100+ languages on every modality
    
-   Open Weight
    
    Open weight option available
    

Workflows

## One model you trust instead of a shaky pipeline

Documents and audio in, structured output back. No OCR tool, speech API, classifier, LLMs and glue code to wire together.

[Get started](https://interfaze.ai/docs)

Models

## Two models, one API

Interfaze 1 for the highest accuracy, Interfaze 1.1 Lite with open weights you can self-host.

[View models](https://interfaze.ai/docs/models)

### [Interfaze 1](https://interfaze.ai/docs/models/interfaze-1)

Our flagship model for the highest accuracy on deterministic tasks.

Input

$1.50

/ MTok

Output

$3.50

/ MTok

Context window

1M tokens

Max output

32k tokens

Reasoning

Available

Input

Text · Images · Audio · File · Video

### [Interfaze 1.1 Lite](https://interfaze.ai/docs/models/interfaze-1.1-lite)

[Open weights](https://huggingface.co/interfaze-ai/interfaze-1-lite)

Smaller open weight model. Available for self-hosting.

Input

$0.85

/ MTok

Output

$1.50

/ MTok

Context window

128k tokens

Max output

32k tokens

Reasoning

Available

Input

Text · Images · Audio · File

Developers

## Works with the SDK you already use

Official TypeScript and Python SDKs, and OpenAI compatible, so the OpenAI SDK, Vercel AI SDK and LangChain work unchanged.

Setup with agents

**Interfaze SDK · typescript**

```typescript
import { Interfaze, responseFormat } from "interfaze";
import { z } from "zod";

const interfaze = new Interfaze({
    apiKey: "<your-api-key>"
});

const IDSchema = z.object({
    first_name: z.string().describe("First name on the ID"),
    last_name: z.string().describe("Last name on the ID"),
    dob: z.string().describe("Date of birth on the ID"),
    driver_licence_number: z.string().describe("Driver licence number on the ID"),
});

const response = await interfaze.chat.completions.create({
    messages: [
        {
            role: "user",
            content: [
                { type: "text", text: "Extract the details from this ID" },
                {
                    type: "image_url",
                    image_url: {
                        url: "https://r2public.jigsawstack.com/interfaze/examples/id.jpg",
                    },
                },
            ],
        },
    ],
    response_format: responseFormat(z.toJSONSchema(IDSchema), "id_schema"),
});

console.log(JSON.parse(response.choices[0]?.message.content ?? "{}"));
```

**Vercel AI SDK · typescript**

```typescript
import { createInterfaze } from "@interfaze-ai/ai-sdk";
import { generateText, Output } from "ai";
import { z } from "zod";

const interfaze = createInterfaze({
    apiKey: "<your-api-key>"
});

const IDSchema = z.object({
    first_name: z.string().describe("First name on the ID"),
    last_name: z.string().describe("Last name on the ID"),
    dob: z.string().describe("Date of birth on the ID"),
    driver_licence_number: z.string().describe("Driver licence number on the ID"),
});

const { output } = await generateText({
    model: interfaze("interfaze-1"),
    output: Output.object({ schema: IDSchema }),
    messages: [
        {
            role: "user",
            content: [
                { type: "text", text: "Extract the details from this ID" },
                {
                    type: "image",
                    mediaType: "image/jpeg",
                    image: "https://r2public.jigsawstack.com/interfaze/examples/id.jpg",
                },
            ],
        },
    ],
});

console.log(output);
```

**LangChain SDK · typescript**

```typescript
import { ChatInterfaze } from "@interfaze-ai/langchain";
import { z } from "zod";

const interfaze = new ChatInterfaze({
    apiKey: "<your-api-key>",
});

const IDSchema = z.object({
    first_name: z.string().describe("First name on the ID"),
    last_name: z.string().describe("Last name on the ID"),
    dob: z.string().describe("Date of birth on the ID"),
    driver_licence_number: z.string().describe("Driver licence number on the ID"),
});

const structuredModel = interfaze.withStructuredOutput(IDSchema);

const response = await structuredModel.invoke([
    {
        role: "user",
        content: [
            { type: "text", text: "Extract the details from this ID" },
            {
                type: "image_url",
                image_url: {
                    url: "https://r2public.jigsawstack.com/interfaze/examples/id.jpg",
                },
            },
        ],
    },
]);

console.log(response);
```

**Interfaze SDK · python**

```python
from interfaze import Interfaze
from pydantic import BaseModel, Field

interfaze = Interfaze(
    api_key="<your-api-key>"
)

class IDSchema(BaseModel):
    first_name: str = Field(..., description="First name on the ID")
    last_name: str = Field(..., description="Last name on the ID")
    dob: str = Field(..., description="Date of birth on the ID")
    driver_licence_number: str = Field(..., description="Driver licence number on the ID")

response = interfaze.chat.completions.parse(
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "Extract the details from this ID"},
                {
                    "type": "image_url",
                    "image_url": {
                        "url": "https://r2public.jigsawstack.com/interfaze/examples/id.jpg"
                    },
                },
            ],
        }
    ],
    response_format=IDSchema,
)

print(response.choices[0].message.parsed)
```

**LangChain SDK · python**

```python
from interfaze_langchain import ChatInterfaze
from langchain_core.messages import HumanMessage
from pydantic import BaseModel, Field

interfaze = ChatInterfaze(api_key="<your-api-key>")

class IDSchema(BaseModel):
    first_name: str = Field(..., description="First name on the ID")
    last_name: str = Field(..., description="Last name on the ID")
    dob: str = Field(..., description="Date of birth on the ID")
    driver_licence_number: str = Field(..., description="Driver licence number on the ID")

structured_llm = interfaze.with_structured_output(IDSchema)

response = structured_llm.invoke([
    HumanMessage(
        content=[
            {"type": "text", "text": "Extract the details from this ID"},
            {
                "type": "image_url",
                "image_url": {
                    "url": "https://r2public.jigsawstack.com/interfaze/examples/id.jpg"
                },
            },
        ]
    )
])

print(response)
```

Benchmarks

## How it compares

Public benchmarks against models in the same price tier.

[Full breakdown](https://interfaze.ai/leaderboards)

Overview

OCRBench V2

olmOCR

RefCOCO

VoxPopuli-Cleaned-AA

SOB Value Acc

Spider-2.0-Lite

GPQA Diamond

MMMLU

MMMU-Pro

Radar chart

| Benchmark | Interfaze-1 | Interfaze-1.1-Lite | Gemini-3.8-Flash | Claude-Sonnet-5 | GPT-5.4-Mini | Grok-4.3 |
| --- | --- | --- | --- | --- | --- | --- |
| OCRBench V2 Native OCR | 70.7% | 60.9% | 63.9% | 59.2% | 52.7% | 54.7% |
| olmOCR Complex document processing | 85.7% | 83.8% | 82.3% | 83.5% | 80.1% | 81.9% |
| RefCOCO Object detection (NL prompts) | 82.1% | 83.8% | 80.9% | 69.2% | 67.0% | 25.0% |
| VoxPopuli-Cleaned-AA ASR (speech recognition) | 2.4% | 3.0% | 4.0% | — | — | — |
| SOB Value Acc Structured output | 80.5% | 81.5% | 80.2% | 79.3% | 75.1% | 78.4% |
| Spider-2.0-Lite Text-to-SQL | 52.9% | 48.9% | 46.7% | 50.6% | 26.7% | 45.9% |
| GPQA Diamond PhD-level problem solving | 92.4% | 85.9% | 91.4% | 78.3% | 82.8% | 73.6% |
| MMMLU Multilingual Q&A | 90.9% | 87.8% | 88.1% | 87.3% | 75.3% | 89.7% |
| MMMU-Pro Multimodal understanding | 71.1% | 73.2% | 64.2% | 53.3% | 40.4% | 68.7% |

\*Bold cells mark the leader on each benchmark. '—' means the model doesn't support that modality.

Verifiable output

## Every value comes back with its box and confidence

Check any field against the page, and send low-confidence values to review instead of trusting them blindly.

[OCR docs](https://interfaze.ai/docs/vision/ocr)

Structured output

## Type safe output, zero hallucinations

Bring your own JSON schema. Every response comes back with the consistent output structured, confidence scores, metadata like bounding boxes that are machine safe.

[Structured output docs](https://interfaze.ai/docs/structured-output)

Multilingual

## 100+ languages, no translation step

Text, audio and images in over 100 languages, read and translated in the same call.

[Translation docs](https://interfaze.ai/docs/translation)

Architecture

## Specialised task-specific models, one transformer

A Mixture-of-Architecture model (MoA): task-specific DNNs joined by a transformer trained for each task. The precision of specialized models with the flexibility of an LLM.

[Read paper](https://www.arxiv.org/abs/2602.04101)

Deployment

## Runs anywhere

[Talk to sales](https://interfaze.ai/sales)

### [Interfaze API](https://interfaze.ai/pricing)

Get an API key and start calling. Turn on zero data retention per request.

Region

Global availability

Data

Zero data retention header

Hardware

Hosted by us

Pricing

Pay per token

### [Dedicated cloud](https://interfaze.ai/sales)

Your own Interfaze deployment in our cloud, with volume pricing, SLAs and a private support channel.

Region

Region lock available

Data

SOC2 & HIPAA agreements

Hardware

Dedicated GPUs, hosted by us

Pricing

Volume, with SLAs

### [Your VPC](https://interfaze.ai/sales)

Interfaze 1 deployed inside your own cloud, so data never leaves your network.

Region

Your cloud's region

Data

Stays in your network

Hardware

Your GPUs

Pricing

Custom

### [Open weights](https://interfaze.ai/sales)

[Hugging Face](https://huggingface.co/interfaze-ai/interfaze-1-lite)

Download Interfaze 1.1 Lite from Hugging Face and run it on your own hardware either on-prem & air-gapped.

Region

Wherever you run it

Data

Stays on your machine

Hardware

Your GPUs

Pricing

Custom

Blog

## Latest blog posts

[All blogs](https://interfaze.ai/blog)

-   [Interfaze 1.1 Lite: The Truly Open Weight Model for Deterministic Work](https://interfaze.ai/blog/introducing-interfaze-1-1-lite)
-   [The First Open Weight Model for Deterministic Work: Interfaze 1 Lite](https://interfaze.ai/blog/the-first-open-weight-model-for-deterministic-work-interfaze-1-lite)
-   [Jev, now open source: Lev](https://interfaze.ai/blog/jev-now-open-source-lev)

## FAQs

[All FAQs](https://interfaze.ai/docs/faqs)

What are the rate limits?

You can make 50 requests per second. If you need more, please contact us at support@interfaze.ai

Are my messages/prompts stored?

Request logging is on by default, and you can view your prompts and responses in the dashboard. To disable storage of request contents, enable Zero Data Retention with the x-interfaze-zdr: true header. Request metadata, including token counts, is still kept.

How do I report bugs?

We do expect many bugs and issues. Please report them to us at support@interfaze.ai

How do I get in contact with the team for other inquiries?

Please email us at support@interfaze.ai or reach out to Yoeven on X: https://x.com/yoeven

Do you charge for infrastructure usage like Sandbox containers, Browser engine, etc.?

We don't have a separate charge for infrastructure usage. Infrastructure usage is converted to tokens and added to both input and output tokens. For example, for a code execution, we'll add input tokens for the code inputted and output tokens for the output generated by the sandbox regardless for how long it took to execute.
