Developer API

Tacita's API is OpenAI-compatible. Point any OpenAI SDK or tool at https://api.tacita.chat/v1, use a Tacita API key, and your existing code works: chat completions, streaming, images and video.

• Wallet-owned keys. Create keys on tacita.chat after signing in with your wallet. There is no email and no account to set up. See API keys & limits.
• Pay for what you use. Usage is billed to the wallet that owns the key. Free daily credits go first if you hold the token, then paid credits. See Billing.
• Anonymous mode. Requests run on a secondary compute market of trusted providers. They see the prompt but never who sent it. TEE mode through the API is not available yet. See Models & modes.

Quickstart#

1. Create a key#

Open tacita.chat → Developers, sign in with your wallet and click Create key. The key looks like tc_live_… and is shown once, so store it somewhere safe.

export TACITA_API_KEY="tc_live_..."

2. Call it with the OpenAI SDK#

JavaScript / TypeScript

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.TACITA_API_KEY,
  baseURL: "https://api.tacita.chat/v1",
});

const r = await client.chat.completions.create({
  model: "deepseek-v4.1-flash",
  messages: [{ role: "user", content: "Write a haiku about silence." }],
});
console.log(r.choices[0].message.content);

Python

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["TACITA_API_KEY"], base_url="https://api.tacita.chat/v1")

r = client.chat.completions.create(
    model="claude-sonnet-5.5",
    messages=[{"role": "user", "content": "Explain zero-knowledge proofs in two sentences."}],
)
print(r.choices[0].message.content)

curl

curl https://api.tacita.chat/v1/chat/completions \
  -H "Authorization: Bearer $TACITA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-4o-mini","messages":[{"role":"user","content":"Hello!"}]}'

3. Stream#

const stream = await client.chat.completions.create({
  model: "claude-sonnet-5.5",
  stream: true,
  messages: [{ role: "user", content: "Tell me a short story." }],
});
for await (const chunk of stream) process.stdout.write(chunk.choices[0]?.delta?.content ?? "");

The last chunk carries usage (prompt and completion tokens).

Images and video#

const img = await client.images.generate({ model: "seedream-4.5", prompt: "a lighthouse at dusk, oil painting", size: "1024x1024" });
console.log(img.data[0].url); // temporary link, valid 10–30 minutes

const job = await client.videos.create({ model: "ltx-2-fast-text-to-video", prompt: "a paper boat on a rainy street", seconds: "6" });
// poll client.videos.retrieve(job.id) until status is "completed", then client.videos.downloadContent(job.id)

Image and video prompts go through a safety check first. Clearly illegal content is refused. Video is in beta.

Safe retries for images. Send your own id in the X-Tacita-Request-Id header (8–64 characters of letters, digits, _ or -), for example client.images.generate(params, { headers: { "X-Tacita-Request-Id": id } }). If the connection drops, GET /media/by-request/{id} returns the images, and repeating the request with the same id never generates or charges twice. An image is charged when it is delivered; one that is never delivered is refunded automatically.

Endpoints#

Privacy#

• Keys are stored only as a hash. We keep no request logs, no prompts, no responses and no IP addresses.
• Per key we keep only a name, the creation date, an optional spend limit and this month's spend counter.
• Generated media is scrubbed of GPS and device metadata, and AI-provenance labels are kept.

For more, see Data handling.

Next#

• API keys & limits
• Billing & credits
• Models & modes
• Errors & rate limits
• @tacita/sdk