ChatGPT API without code: text, JSON and images in your automations

You can use the ChatGPT API without code in nana800 with one block: paste your OpenAI key, pick a model and write the task, filling it with data from a lead, an email or a Telegram message. The block answers in text or JSON, reads images by link and generates new ones. Each block run costs $0.002 in nana800, and OpenAI bills the model tokens to your own OpenAI account at its prices.

Build it in the builder
  • Block run: $0.002
  • Free every month: $3.50
  • No card needed

What you can do

  • Reply to customers in a Telegram bot, by email or on your website with text written by ChatGPT
  • Summarize emails and leads, check the tone of a review, sort requests by topic
  • Get the answer as JSON and put its fields into table cells or CRM fields
  • Describe a photo by link or compare two images: a product, a receipt, a screenshot, a document
  • Generate an image from a description or edit a reference image
  • Send requests through an intermediary with the same request format by setting its API address

Blocks in the builder

  • OpenAI request

    Action · $0.002 per run

    Sends a task to the chosen GPT model with your key. Four modes: answer a text prompt, analyze one image, compare two images and generate an image. The result comes back as text, JSON or a JPEG image.

How to set it up

  1. 1

    Get an OpenAI API key

    In the OpenAI developer dashboard open API keys and click Create new secret key. The key starts with sk- followed by a long string. Your OpenAI account needs a positive balance, otherwise OpenAI rejects the requests.

  2. 2

    Create a scheme and add a trigger

    Create a scheme in the builder and choose what starts it: webhook for website forms, Telegram message from user to bot for a bot, or button to run it by hand. The trigger passes on the text to process.

  3. 3

    Add the Open AI request block

    Pick a mode such as TEXT_COMPLETION, paste the key, choose a model and write the prompt with data from the trigger. Use the system instruction to set the role and tone, for example: you are a polite support agent, keep answers short.

  4. 4

    Choose the answer format

    Keep text for plain answers. To split the answer into fields, choose json_object and list the fields you need in the prompt: the parsed object lands in json_response, and the next blocks can read single values from it.

  5. 5

    Pass the answer on

    Add a block that delivers the result: Telegram message send, send email message or google sheet record. The answer text is in text_response. In IMAGE_GENERATION mode the image comes in the base64 field, and the save file block stores it.

  6. 6

    Turn the scheme on and test it

    Turn the scheme on, press the button trigger and check the model answer in the run history of the scheme. The button itself is free. If the answer looks right, keep the scheme on. While the scheme is on, it costs $0.0042 per hour.

Popular connections

  • OpenAI (ChatGPT) + Telegram

    ChatGPT writes posts and a bot publishes them to your Telegram channel on the days and hours you choose, with no code. Free balance every month, no card needed.

  • Telegram + OpenAI (ChatGPT)

    Build a ChatGPT bot for Telegram without code: it answers messages through your own OpenAI API key. Free balance every month, no card needed.

  • Webhook + OpenAI (ChatGPT)

    Get your own API address for ChatGPT without code: your site sends a request, an OpenAI model answers with text or JSON. Free balance every month, no card needed.

Frequently asked questions

A block run costs $0.002 whatever the answer length. OpenAI counts the model tokens separately and charges them to your OpenAI account. An enabled scheme costs $0.0042 per hour. Your nana800 balance is topped up to $3.50 for free every month, no card needed. A scheme kept on around the clock for a month spends most of that on scheme hours.

In the OpenAI developer dashboard: API keys, then Create new secret key. Paste it into the OpenAI API key field, or keep it in the project variables and insert it from there. Your OpenAI account needs a positive balance for requests to work.

The default gpt-6.1-sol fits most tasks. gpt-6-astra is the smartest, gpt-6-luna is fast and cheap. Models at the end of the list, such as gpt-4o and gpt-4.1, are being retired by OpenAI, and gpt-4.1-nano and o4-mini no longer work.

Yes. Put the address of a service that accepts requests in the OpenAI format into the API address field. When the field is empty, requests go straight to OpenAI. Addresses of internal networks are rejected.

The gpt-5, gpt-6 and o3 models reason before answering, and that reasoning counts toward the answer length limit. If you leave the limit empty, the block sets 25000 tokens. If the limit runs out, finish_reason is length and the block fails with a clear reason.

Choose IMAGE_GENERATION and describe the picture in the prompt. Sizes: auto, square 1024x1024, portrait 1024x1536 or landscape 1536x1024. Quality: auto, low, medium or high. To edit an existing picture, give a link to a reference image. The result is a JPEG.

Build your first scheme for free

Sign-up is free and needs no card. Every month the balance is topped up to $3.50 for free, enough for about 3,500 runs of basic blocks.

Build it in the builder