Quickstart

Make your first API call in under five minutes: create an account, issue a key, and send a chat completion.

Updated Sep 2026

This guide walks through the fastest path from signalling an empty account to receiving a production response from the PARROTSIGHT API. By the end you will have a working key, a completed request to ps-lingua-7b, and a clear sense of where to go next.

Everything below targets the public REST base URL, https://api.parrotsight.com/v1. Token-based LLM calls are also served from the same base URL through an OpenAI-compatible surface, so tooling you already use can be pointed at us with a single line change.

1. Create an account

Start at the signup page and register with a company email. Accounts are provisioned immediately, and the Developer tier requires no credit card and no commitment.

1,000 free API calls every month, refreshed on the first day of the billing cycle.
Access to all six production models, including the ps-lingua-7b beta.
Live and test keys, usage dashboards, and per-key scopes from day one.

2. Issue an API key

Open the Console, navigate to API Keys, and choose Create key. Name it after the environment it will serve, then copy the secret immediately — it is shown once. Store the value in an environment variable rather than committing it to source control.

terminal

export PS_API_KEY="sk-ps-live-9f2c41d8a1b04e7b9c3e5d6f7a8b9c0d"

3. Make your first request

The chat completions endpoint accepts an OpenAI-style payload. The request below asks ps-lingua-7b to condense a concept into a single pair of sentences, with temperature held low for deterministic output.

request.sh

curl https://api.parrotsight.com/v1/chat/completions \
  -H "Authorization: Bearer $PS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ps-lingua-7b",
    "messages": [
      { "role": "system", "content": "You are a concise technical assistant." },
      { "role": "user", "content": "Explain zero-shot classification in two sentences." }
    ],
    "temperature": 0.2,
    "max_tokens": 128,
    "stream": false
  }'

4. Read the response

A successful request returns the generated message together with token accounting. The usage block lets you reconcile spend against the per-1M-token rate shown on the pricing page.

response.json

{
  "id": "chatcmpl-9f2c41d8a1b04e7b9c3e5d6f7a8b9c0d",
  "object": "chat.completion",
  "created": 1756684800,
  "model": "ps-lingua-7b",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Zero-shot classification assigns a label to input text without any task-specific training examples, by comparing the input against candidate labels using general language semantics the model already knows."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 31,
    "completion_tokens": 42,
    "total_tokens": 73
  }
}

Next steps

You can now call the API. The following documents cover the pieces most teams need before promoting an integration to production.

Authentication and key lifecycle — live vs test keys, rotation, and IP allowlisting.
API Reference — the complete endpoint catalogue with request fields and response shapes.
Rate limits — how plan limits, concurrent requests and monthly quotas interact.
SDKs & Libraries — the first-party Python and Node clients, plus OpenAI-compatible usage.

Was this helpful?

Your feedback shapes how we improve this documentation.

YesNo