Remove Text API

A production Remove Text API that finds watermarks, date stamps, captions and other lettering on a photo and fills the pixels underneath. One HTTP call. $0.005 per image (5 credits). 5,000 free credits (24-hour trial), no credit card.

Quick start — Remove Text API

Upload the image. You do not have to box the watermark. The Remove Text API detects lettering and reconstructs the pixels underneath. Use optional prompt when only one mark should go.

curl -X POST https://api.pixelapi.dev/v1/image/remove-text \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "image=@photo.jpg" \
  -F "languages=en"

# Optional: clamp to one mark
# -F 'prompt=remove the "ACME" watermark'

curl https://api.pixelapi.dev/v1/image/GENERATION_ID \
  -H "Authorization: Bearer YOUR_API_KEY"

import requests
r = requests.post(
    "https://api.pixelapi.dev/v1/image/remove-text",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("photo.jpg", "rb")},
    data={"languages": "en", "prompt": 'remove the "ACME" watermark'},
)
job = r.json()

import fs from "fs";
const form = new FormData();
form.append("image", fs.createReadStream("photo.jpg"));
form.append("languages", "en");
const res = await fetch("https://api.pixelapi.dev/v1/image/remove-text", {
  method: "POST",
  headers: { Authorization: "Bearer " + process.env.PIXELAPI_KEY },
  body: form,
});

package main
// POST multipart field "image" to https://api.pixelapi.dev/v1/image/remove-text
// Header: Authorization: Bearer YOUR_API_KEY
// Then GET /v1/image/{id} until status == "completed"

What the Remove Text API is for

Stock sites, screenshot pipelines, and user-upload apps collect frames that already have a date in the corner, a social handle, or a translucent agency watermark. Brushing those out by hand does not scale. This Remove Text API is a dedicated REST route: it finds glyphs, builds a tight mask, and fills the region so the photo can ship.

It is cheaper and stricter than general object removal. Live price is $0.005 per image (5 credits). Do not quote the object-removal rate (12.5 credits) and do not quote retired numbers from old docs. If a client asks to delete a person or a cable, send them to the Remove Object API. If they want a transparent cutout, use Background Removal.

Pricing comparison

5,000 free credits (24-hour trial), no card. After the trial, prepaid credits. Always read /v1/pricing/live in production so a dashboard copy cannot drift from the meter.

Production notes

• Multipart POST. Fields: image (required), languages (default en), prompt (optional target).
• Poll GET /v1/image/{id}. Download output_url only when status is completed.
• Over-removal is the failure mode that matters: a good job leaves packaging type and street signs that belong in the photo. Name the watermark when the frame has both.
• Do not send this endpoint a job whose only problem is a person in the background. The detector will look for glyphs and leave the person.
• Web UI: remove text / watermark tool if you want to try a frame before you wire the key.

Use cases and related pages

Newsrooms strip agency bugs from wires. Marketplaces strip seller timestamps. Photo apps strip the camera-clock overlay grandparents still enable. Screenshot products strip UI chrome text that leaked into a capture.

Related how-tos live under object-removal because many teams start there by mistake: remove text, remove watermark. After the fill, print pipelines often call the Upscale API. Semantic restyle (not glyph erase) is the AI Image Editor API. Hubs: use cases, alternatives, howto, docs.

Frequently asked questions