EvaEsi Instant | EvaEsi Dev

Overview

Instant is the default user-facing model. Use it for general questions, short factual lookups, summaries, and conversational tasks. Bills usage in rungs (one rung = one byte processed), so request size and billing stay aligned.

When you ask for live evidence via web_search, retrieved sources come back on the response payload with an evidence_verdict. If retrieved sources disagree, Instant declines to take a side rather than guess.

Endpoint

POST https://evaesi.dev/api/v1/responses

Pass "model": "instant" in the JSON body, or omit model entirely, Instant is the default.

Capabilities

General-purpose chat and analysis
Server-sent event streaming via stream: true
Live evidence retrieval via web_search: true
Citations returned on the response with stance and source-tier metadata
Same-origin sources lookup via sources_url while the request stays resident

Batch and tool-calling are not yet exposed for any model. Treat both as unavailable until they land in the public release notes.

Limits

Field	Value
Max input rungs	`100,000`
Max output rungs	`32,000`
Context window	`4,096` model bytes (billing rungs are a separate unit)
Streaming	Yes
Batch	No. The route is mounted but returns `503 EVA_UNAVAILABLE`
Tools	No

Prompt validation enforces a 100,000-byte hard cap on the prompt field at the API edge; over-size prompts return 400 PROMPT_TOO_LARGE. Empty prompts return 400 EMPTY_PROMPT.

Billing

Instant runs at the base rate, multiplier 1.0×. Input and output rungs are metered separately and reported on both the response body (usage object) and headers.

Headers on a successful non-stream response:

x-evaesi-input-rungs: 29
x-evaesi-output-rungs: 118
x-evaesi-total-rungs: 147
x-ratelimit-limit: 300
x-ratelimit-remaining: 287

The request

Field	Type	Required	Description
`prompt`	string	Yes	Prompt text. Max 100,000 bytes.
`model`	string?	No	Optional. Defaults to `instant`.
`max_output_rungs`	u64?	No	Output cap, validated against the model's 32,000 limit (`400 INVALID_REQUEST` above it). Falls back to the model output limit.
`web_search`	bool	No	Request live retrieval and citations.
`stream`	bool?	No	Server-sent event streaming on the same route.

Response

{
  "request_id": "req_8f3a...",
  "content": "Generated answer...",
  "sources": [],
  "sources_url": "/api/v1/sources/req_8f3a...",
  "evidence_verdict": null,
  "usage": {
    "input_rungs": 29,
    "output_rungs": 118,
    "total_rungs": 147
  },
  "model": "instant"
}

sources_url returns the same-origin path /api/v1/sources/{request_id} while the gateway still holds the live evidence bundle.

Examples

cURL

curl -X POST https://evaesi.dev/api/v1/responses \
    -H "Authorization: Bearer esi_dev_YOUR_KEY" \
    -H "Content-Type: application/json" \
  -d '{
    "prompt": "Summarise the news from the last 24 hours.",
    "model": "instant",
    "web_search": true,
    "max_output_rungs": 1024
  }'

Python

import requests

response = requests.post(
    "https://evaesi.dev/api/v1/responses",
    headers={"Authorization": "Bearer esi_dev_YOUR_KEY"},
    json={
        "prompt": "Summarise the news from the last 24 hours.",
        "model": "instant",
        "web_search": True,
        "max_output_rungs": 1024,
    },
)

data = response.json()
print(data["content"])
print(f"sources: {len(data['sources'])}")

JavaScript

const response = await fetch("https://evaesi.dev/api/v1/responses", {
  method: "POST",
  headers: {
    "Authorization": "Bearer esi_dev_YOUR_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    prompt: "Summarise the news from the last 24 hours.",
    model: "instant",
    web_search: true,
  }),
});

const data = await response.json();
console.log(data.content);

Start from a live reference page.