Cehpoint AI API

Endpoint

Send requests to this URL. No authentication headers needed. Just Content-Type: application/json and your messages.

Model

Agent Tools (Function Calling)

Cehpoint AI is not just a text model — it is an agent. When your prompt needs a live fact, a calculation, a lookup or a page read, it automatically calls the right built-in tool and answers using the real result. This works on both the free and commercial endpoints — no setup, no extra parameters. Just ask naturally; the model decides when to use a tool.

Vision — Send Images

Cehpoint AI accepts images alongside text, so it can read documents, extract data from screenshots, describe photos, or analyse diagrams. Use the standard OpenAI-compatible multimodal message format — an image_url content part (a public URL or a data: base64 URI):

curl https://ai.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [{ "role": "user", "content": [
      { "type": "text", "text": "Extract the invoice number and total from this image." },
      { "type": "image_url", "image_url": { "url": "https://example.com/invoice.jpg" } }
    ]}]
  }'

Commercial keys use the same format on the /capi/v1/chat/completions endpoint. Both hosted images (public URLs) and inline base64 data: URIs are supported.

Request Parameters

The only two things you must send are model and messages. Everything else is optional.

Message Roles

Use system to control personality/tone. Use assistant for follow-up conversations.

Response Format

{
  // standard chat completion response
  "model": "cehpoint-ai",
  "provider": "ai-api.cehpoint.co.in",
  "choices": [{
    "message": { "content": "Hello!" },
    "finish_reason": "stop"
  }],
  "usage": { "total_tokens": 52 }
}

Standard JSON response structure. The provider field is set to our domain. model is always cehpoint-ai.

Error Codes

Limits

Text-to-Speech

Convert text to natural speech. Premium accounts get Sarvam AI's Bulbul HD neural voice; free and anonymous requests use the standard voice. Returns binary audio (audio/wav or audio/mpeg).

curl https://ai-api.cehpoint.co.in/v1/tts \
  -H "Content-Type: application/json" \
  -d '{"text":"Hello from Cehpoint AI","lang":"en","gender":"female"}' \
  --output speech.wav

Image Generation

Generate an image from a text prompt. The call is synchronous — images typically return in under ten seconds — and the response is OpenAI-compatible, so an existing SDK can be pointed at Cehpoint without code changes. Assets are served from Cehpoint infrastructure.

curl https://ai-api.cehpoint.co.in/capi/v1/images/generations \
  -H "Authorization: Bearer $CEHPOINT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "a red ceramic teacup on a wooden table, soft daylight, product photo",
    "aspect_ratio": "16:9",
    "size": "2K"
  }'

{
  "created": 1786326431,
  "size": "2K",
  "aspect_ratio": "16:9",
  "data": [
    { "url": "https://ai-api.cehpoint.co.in/v1/media/file/a152d9049d0ce8262999.png" }
  ]
}

An unrecognised value is replaced with the default rather than rejected, so a request never fails because of a shape you cannot use. The response always states the size and aspect_ratio actually used — read those rather than assuming your request was applied verbatim.

Notes. The returned URL is permanent for the retention window configured on your plan and may be fetched without authentication, so it can be embedded directly. Prompts are limited to 1200 characters.

Video Generation

Generate a short video from a text prompt. Video takes appreciably longer than an image, so the request returns immediately with a job id and you poll for completion. The job is durable — it survives client disconnects, so you can submit now and collect the result later.

curl https://ai-api.cehpoint.co.in/capi/v1/videos \
  -H "Authorization: Bearer $CEHPOINT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "a slow cinematic pan over calm ocean water at sunrise",
    "aspect_ratio": "16:9",
    "resolution": "1080p",
    "duration_seconds": 10
  }'

{ "id": "ddafad9dc549f51264da", "object": "video", "status": "queued", "created": 1786326487 }

Frame sizes are standard for each combination — for example 1080p at 16:9 is 1920×1080, and at 9:16 is 1080×1920. The completed job reports resolution, aspect_ratio, duration_seconds and the exact pixel size delivered. Longer or higher-resolution videos take proportionally longer to render.

Poll until status is completed:

curl https://ai-api.cehpoint.co.in/capi/v1/videos/ddafad9dc549f51264da \
  -H "Authorization: Bearer $CEHPOINT_API_KEY"

{
  "id": "ddafad9dc549f51264da",
  "object": "video",
  "status": "completed",
  "url": "https://ai-api.cehpoint.co.in/v1/media/file/ddafad9dc549f51264da.mp4",
  "error": null
}

Polling guidance. Poll no more often than every five seconds. Most clips complete in one to two minutes; treat anything beyond twenty minutes as failed and resubmit.

Supported Shapes and Sizes

Rather than hard-coding the lists above, you can read them at run time. This endpoint is the same source our own interface uses, so it is always current.

curl https://ai-api.cehpoint.co.in/capi/v1/media/options \
  -H "Authorization: Bearer $CEHPOINT_API_KEY"

{
  "ok": true,
  "options": {
    "image": {
      "aspect_ratios": ["1:1","4:3","3:4","16:9","9:16","3:2","2:3","21:9"],
      "sizes": ["1K","2K","3K","4K"],
      "defaults": { "size": "2K", "aspect_ratio": "1:1" }
    },
    "video": {
      "aspect_ratios": ["16:9","9:16","1:1","4:3","3:4"],
      "resolutions": ["480p","720p","1080p"],
      "duration_options": [5, 10],
      "duration_seconds": { "options": [5, 10], "default": 5, "fps": 24 },
      "defaults": { "resolution": "720p", "aspect_ratio": "16:9", "duration_seconds": 5 }
    }
  }
}

Cehpoint Reader

Cehpoint Reader turns any web page or document URL into clean, readable text for LLM and RAG pipelines — it fetches the page, strips scripts, navigation and ads, and returns the main content, title and links as JSON. It runs on its own base URL and does not require a Cehpoint AI key.

For a higher daily allowance, create a Reader key (it begins with cehr_) from your dashboard and send it as Authorization: Bearer cehr_your_key.

curl https://r.cehpoint.co.in/ \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'

Cehpoint Code

Cehpoint Code is an AI engineer that turns a plain-language request into a finished, downloadable deliverable — a website or tool, a business document or model, a data analysis, or a report. Each build runs privately in an isolated sandbox and returns the produced files. Use it in the browser at /code, or via the API below with your Cehpoint AI key (cehp_) or a Commercial key (cehk_ — dedicated quota).

The 5-minute policy. Every build has a strict five-minute limit so the queue stays fast. If a build needs longer it returns status: "timeout" — re-submit to run it again as a fresh build (which counts as your next build). Free keys get a daily allowance; Commercial keys get a dedicated, much higher quota.

curl https://ai.cehpoint.co.in/v1/code/run \
  -H "Authorization: Bearer cehp_your_key" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Build a landing page for a bakery", "type": "web"}'

Team & Invites

A commercial licence owner can bring their team onto the same account. Invites are managed from the dashboard and via these session-authenticated endpoints.

Roles. Owner — full control of keys, grounding and team. Member — can use the API and manage the knowledge base, but not keys, grounding or team. Raw invite tokens are never stored (only a hash) and never logged.

Commercial API

The public endpoint above is free and key-less for builders and prototypes. When you ship to production — with commercial-use rights, higher throughput, dedicated support, GST invoicing and your own licence key — move to the Commercial API. It's the same OpenAI-compatible interface and the same cehpoint-ai model, just an authenticated, metered base URL. See plans & get a key →

Authentication

Unlike the free endpoint, every commercial request is authenticated with your licence key as a Bearer token:

Keys begin with cehk_ and are issued from your dashboard once a licence is active. Treat a key like a password — keep it server-side. Requests without a valid key return 401 Unauthorized.

Endpoint & migration

Same request and response shape as the free API — migrating is two lines: point base_url at /capi/v1 and add your key. Any OpenAI SDK or existing OpenAI integration works unchanged.

# Drop-in: same OpenAI SDK, two lines changed
from openai import OpenAI

client = OpenAI(
  base_url="https://ai-api.cehpoint.co.in/capi/v1",
  api_key="cehk_your_licence_key",
)

resp = client.chat.completions.create(
  model="cehpoint-ai",
  messages=[{"role": "user", "content": "Hello"}],
)
# set stream=True for SSE, exactly like the free endpoint

Plans & limits

Pricing is usage-sized — tell us your expected volume and our AI sizes a plan and quotes you in minutes. Every commercial plan includes commercial-use rights, dedicated support and a GST invoice for India.

Use Cases

Every example below was tested live against the API. Pick a category, copy the code, and run it.

curl https://ai-api.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [{"role": "user", "content": "Explain quantum computing in 3 sentences."}]
  }'

import requests

resp = requests.post(
    "https://ai-api.cehpoint.co.in/v1/chat/completions",
    json={
        "model": "cehpoint-ai",
        "messages": [{"role": "user", "content": "Explain quantum computing in 3 sentences."}]
    },
)
print(resp.json()["choices"][0]["message"]["content"])

const resp = await fetch("https://ai-api.cehpoint.co.in/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "cehpoint-ai",
    messages: [{ role: "user", content: "Explain quantum computing in 3 sentences." }],
  }),
});
const data = await resp.json();
console.log(data.choices[0].message.content);

from openai import OpenAI

client = OpenAI(
    base_url="https://ai-api.cehpoint.co.in/v1",
    api_key="x",
)

resp = client.chat.completions.create(
    model="cehpoint-ai",
    messages=[{"role": "user", "content": "Explain quantum computing in 3 sentences."}],
)
print(resp.choices[0].message.content)

curl https://ai-api.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [{"role": "user", "content": "Write a Python class for a banking account with deposit, withdraw, and balance check. Include error handling."}]
  }'

import requests

resp = requests.post(
    "https://ai-api.cehpoint.co.in/v1/chat/completions",
    json={
        "model": "cehpoint-ai",
        "messages": [{"role": "user", "content": "Write a Python class for a banking account with deposit, withdraw, and balance check. Include error handling."}]
    },
)
print(resp.json()["choices"][0]["message"]["content"])

const resp = await fetch("https://ai-api.cehpoint.co.in/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "cehpoint-ai",
    messages: [{ role: "user", content: "Write a Python class for a banking account with deposit, withdraw, and balance check. Include error handling." }],
  }),
});
const data = await resp.json();
console.log(data.choices[0].message.content);

from openai import OpenAI

client = OpenAI(
    base_url="https://ai-api.cehpoint.co.in/v1",
    api_key="x",
)

resp = client.chat.completions.create(
    model="cehpoint-ai",
    messages=[{"role": "user", "content": "Write a Python class for a banking account with deposit, withdraw, and balance check. Include error handling."}],
)
print(resp.choices[0].message.content)

curl https://ai-api.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [
      {"role": "system", "content": "You are a friendly support agent."},
      {"role": "user", "content": "My order has not arrived yet."}
    ]
  }'

import requests

resp = requests.post(
    "https://ai-api.cehpoint.co.in/v1/chat/completions",
    json={
        "model": "cehpoint-ai",
        "messages": [
            {"role": "system", "content": "You are a friendly support agent."},
            {"role": "user", "content": "My order has not arrived yet."}
        ]
    },
)
print(resp.json()["choices"][0]["message"]["content"])

const resp = await fetch("https://ai-api.cehpoint.co.in/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "cehpoint-ai",
    messages: [
      { role: "system", content: "You are a friendly support agent." },
      { role: "user", content: "My order has not arrived yet." }
    ],
  }),
});
const data = await resp.json();
console.log(data.choices[0].message.content);

from openai import OpenAI

client = OpenAI(
    base_url="https://ai-api.cehpoint.co.in/v1",
    api_key="x",
)

resp = client.chat.completions.create(
    model="cehpoint-ai",
    messages=[
        {"role": "system", "content": "You are a friendly support agent."},
        {"role": "user", "content": "My order has not arrived yet."}
    ],
)
print(resp.choices[0].message.content)

curl https://ai-api.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [{"role": "user", "content": "Write a professional email apologizing for a delayed delivery and offering a 10% discount."}]
  }'

import requests

resp = requests.post(
    "https://ai-api.cehpoint.co.in/v1/chat/completions",
    json={
        "model": "cehpoint-ai",
        "messages": [{"role": "user", "content": "Write a professional email apologizing for a delayed delivery and offering a 10% discount."}]
    },
)
print(resp.json()["choices"][0]["message"]["content"])

const resp = await fetch("https://ai-api.cehpoint.co.in/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "cehpoint-ai",
    messages: [{ role: "user", content: "Write a professional email apologizing for a delayed delivery and offering a 10% discount." }],
  }),
});
const data = await resp.json();
console.log(data.choices[0].message.content);

from openai import OpenAI

client = OpenAI(
    base_url="https://ai-api.cehpoint.co.in/v1",
    api_key="x",
)

resp = client.chat.completions.create(
    model="cehpoint-ai",
    messages=[{"role": "user", "content": "Write a professional email apologizing for a delayed delivery and offering a 10% discount."}],
)
print(resp.choices[0].message.content)

curl https://ai-api.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [{"role": "user", "content": "Write a 50-word product description for a premium leather wallet."}]
  }'

import requests

resp = requests.post(
    "https://ai-api.cehpoint.co.in/v1/chat/completions",
    json={
        "model": "cehpoint-ai",
        "messages": [{"role": "user", "content": "Write a 50-word product description for a premium leather wallet."}]
    },
)
print(resp.json()["choices"][0]["message"]["content"])

const resp = await fetch("https://ai-api.cehpoint.co.in/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "cehpoint-ai",
    messages: [{ role: "user", content: "Write a 50-word product description for a premium leather wallet." }],
  }),
});
const data = await resp.json();
console.log(data.choices[0].message.content);

from openai import OpenAI

client = OpenAI(
    base_url="https://ai-api.cehpoint.co.in/v1",
    api_key="x",
)

resp = client.chat.completions.create(
    model="cehpoint-ai",
    messages=[{"role": "user", "content": "Write a 50-word product description for a premium leather wallet."}],
)
print(resp.choices[0].message.content)

curl https://ai-api.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [{"role": "user", "content": "Extract names, dates, amounts as JSON: John paid $150 on March 5th, Sarah paid $75 on March 12th."}]
  }'

import requests

resp = requests.post(
    "https://ai-api.cehpoint.co.in/v1/chat/completions",
    json={
        "model": "cehpoint-ai",
        "messages": [{"role": "user", "content": "Extract names, dates, amounts as JSON: John paid $150 on March 5th, Sarah paid $75 on March 12th."}]
    },
)
print(resp.json()["choices"][0]["message"]["content"])

const resp = await fetch("https://ai-api.cehpoint.co.in/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "cehpoint-ai",
    messages: [{ role: "user", content: "Extract names, dates, amounts as JSON: John paid $150 on March 5th, Sarah paid $75 on March 12th." }],
  }),
});
const data = await resp.json();
console.log(data.choices[0].message.content);

from openai import OpenAI

client = OpenAI(
    base_url="https://ai-api.cehpoint.co.in/v1",
    api_key="x",
)

resp = client.chat.completions.create(
    model="cehpoint-ai",
    messages=[{"role": "user", "content": "Extract names, dates, amounts as JSON: John paid $150 on March 5th, Sarah paid $75 on March 12th."}],
)
print(resp.choices[0].message.content)

curl https://ai-api.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [{"role": "user", "content": "List 3 AI applications in healthcare diagnostics."}]
  }'

import requests

resp = requests.post(
    "https://ai-api.cehpoint.co.in/v1/chat/completions",
    json={
        "model": "cehpoint-ai",
        "messages": [{"role": "user", "content": "List 3 AI applications in healthcare diagnostics."}]
    },
)
print(resp.json()["choices"][0]["message"]["content"])

const resp = await fetch("https://ai-api.cehpoint.co.in/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "cehpoint-ai",
    messages: [{ role: "user", content: "List 3 AI applications in healthcare diagnostics." }],
  }),
});
const data = await resp.json();
console.log(data.choices[0].message.content);

from openai import OpenAI

client = OpenAI(
    base_url="https://ai-api.cehpoint.co.in/v1",
    api_key="x",
)

resp = client.chat.completions.create(
    model="cehpoint-ai",
    messages=[{"role": "user", "content": "List 3 AI applications in healthcare diagnostics."}],
)
print(resp.choices[0].message.content)

50 Complete Examples

cURL

curl https://ai-api.cehpoint.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cehpoint-ai",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Python

import requests

resp = requests.post(
    "https://ai-api.cehpoint.co.in/v1/chat/completions",
    json={"model": "cehpoint-ai", "messages": [{"role": "user", "content": "Hello!"}]},
)
print(resp.json()["choices"][0]["message"]["content"])

JavaScript (Fetch)

const resp = await fetch("https://ai-api.cehpoint.co.in/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "cehpoint-ai",
    messages: [{ role: "user", content: "Hello!" }],
  }),
});
const data = await resp.json();
console.log(data.choices[0].message.content);

OpenAI Python SDK

from openai import OpenAI

client = OpenAI(
    base_url="https://ai-api.cehpoint.co.in/v1",
    api_key="x",
)

resp = client.chat.completions.create(
    model="cehpoint-ai",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)