Reads the project itself
Opens files, searches the code and gathers context. You describe the task in plain words, like to a colleague.
Claude, GPT and Gemini write your code. Up to 10× cheaper
Slafy IDE: an editor with an AI agent that reads your project, suggests edits and runs commands. Tokens for Claude, GPT, Gemini and other models below official prices, no subscription.
50,000 tokens on sign-up
Pay only when you use it, no subscriptions
Up to 10× cheaper than official prices
One balance for all models
IDE agent
You set the task, the agent does it in your project
No need to copy code into a chat and back. The agent opens the files it needs, writes the changes and shows them to you before saving to disk. All IDE features and screenshots
01
02
Edits only with your approval
Every change is shown as a diff first. The file changes only after you click “Apply”, and the agent takes rejections into account too.
03
Runs commands
Tests, builds, package installs. Before running, you see the full command text and decide whether to execute it.
04
Any model
Claude, GPT, Gemini, Grok, DeepSeek, Kimi and GLM. Give hard tasks to a strong model and routine to a cheap one.
05
History for each project
Chats are stored separately for each folder and continue after a restart, along with drafts.
06
No prying eyes
Code goes only to the model you choose. The IDE collects no telemetry and doesn't load the editor from the internet.
Browser chat
Just need an answer? Open the chat
After you sign up, the chat opens right away, just like ChatGPT. Nothing to install.
- Switch models in one click right above the conversation
- Under each answer you see how many tokens it used
- All chats on the left, balance and settings in your profile
- Same account and same balance as in the IDE and API
Claude Opus 5.5
Anthropic
Claude Sonnet 5.5
Anthropic
GPT-6.1 Sol
OpenAI
Gemini 4 Argon
DeepSeek V4 Pro
DeepSeek
Pricing
Pay only when you use it. No subscriptions
No monthly fee and no monthly charges. Top up your balance and it's spent only on your requests. Don't use it, don't pay; your balance never expires.
Subscription$0.00
per month. Pay only for what you use
On sign-up50,000
tokens to try, no card or email needed
Most affordable model$0.119
per 1M tokens, DeepSeek V4.1 Flash 1017
Cheaper than official pricesup to 10×
compared to the model developers' rates
$1 gets you about 8.4M tokens of DeepSeek V4.1 Flash 1017, while a subscription to a single AI service costs $20 a month or more.
Claude, GPT, Gemini, Grok and DeepSeek all run on one balance: pick the model for the task and don't overpay for a strong one where a fast one will do. Model errors aren't charged.
API
Your own code, bot or Cursor? Get a key
The API is compatible with OpenAI and Anthropic: change the address and key, and everything else works as before. Create a key in the API tab of your profile.
- Claude Code, Codex, Cursor, Cline, Continue
- OpenAI and Anthropic Python and Node.js SDKs
- Spending limit for each key
curl https://slafy.co/v1/chat/completions \
-H "Authorization: Bearer sk-slafy-your_key" \
-H "Content-Type: application/json" \
-d '{
"model": "claude-sonnet-5-5",
"messages": [{"role": "user", "content": "Hello"}]
}'
Slafy IDE
Download Slafy IDE
Version 0.7.3. After installing, click “Sign in via website”: you sign in through this website, and 50,000 tokens are waiting for new users.
macOS. The build isn't signed by Apple yet. Drag Slafy IDE to Applications and open it. If macOS blocks it, go to System Settings → Privacy & Security and click “Open Anyway”.
Windows. If SmartScreen warns about an unknown publisher, click “More info” → “Run anyway”.
FAQ
How do I pay?
By card in your profile, under “Balance & billing”. Tokens arrive on your balance right after payment, no subscription.
What are tokens and why do models cost differently?
Your balance is counted in base-rate tokens: 1M tokens costs $0.397. Stronger models are more expensive, so each of their tokens uses more balance tokens. Each model's price is listed in pricing, and the exact charge is shown under every answer in the chat.
Why are tokens cheaper than from the model developers?
We buy model access in bulk and sell tokens at a single price for input and output. For many models that's up to 10 times cheaper than the official rate; see the per-model comparison in pricing.
Do tokens expire?
No. Your balance stays on your account until you use it.
How is the IDE different from the chat?
The chat answers questions. The agent in the IDE works with your project files: it reads code, makes edits after you confirm them and runs commands.
Can I change the model's reasoning effort?
Yes. In the chat and the IDE there's a switch next to the input box: “Fast”, “Normal”, “Deep” and a maximum mode. In the IDE it's called “Ultracode”: the model reasons longer, the agent takes more steps and checks its edits with tests.
What do “Goal” and “Check for inconsistencies” do?
The “Check for inconsistencies” button is available in the chat and the IDE: the model rereads its answers and edits and looks for contradictions, errors and unverified claims without changing anything. The IDE has a “Goal” mode and a /goal command: the agent works on its own until it finishes the task and verifies the result with tests or a build.
Is there a referral program?
Yes. Your link is in your profile under “Invite friends”. Every time someone who signed up with it tops up by card, you get 10% of the top-up amount in tokens. More about the program.
How do I contact support?
Message us on Telegram: t.me/Sawasdee_krap. We'll help with payments, signing in, the IDE and API setup.
Can I connect my own app?
Yes. Create a key in the API tab and use the address https://slafy.co/v1. Any client that supports the OpenAI or Anthropic API will work.