Jesse API

One base URL. Your key.

Subscribe, create a key on your account page, and send a chat completion. Anything that speaks the OpenAI API works unchanged.

curl https://jesse.my/api/v1/chat/completions \
  -H "Authorization: Bearer $JESSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "jesse-prod",
    "messages": [{"role": "user", "content": "What is 17 * 23?"}]
  }'

from openai import OpenAI

client = OpenAI(base_url="https://jesse.my/api/v1", api_key=JESSE_API_KEY)

reply = client.chat.completions.create(
    model="jesse-prod",
    messages=[{"role": "user", "content": "What is 17 * 23?"}],
)
print(reply.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://jesse.my/api/v1", apiKey: process.env.JESSE_API_KEY });

const reply = await client.chat.completions.create({
  model: "jesse-prod",
  messages: [{ role: "user", content: "What is 17 * 23?" }],
});
console.log(reply.choices[0].message.content);

Try Jesse

Ask Jesse

Two models. One learns.

Jesse answers the engineering, logic and game tasks it has solvers for. It is not a general-purpose chat model, and it tells you when it cannot answer.

M / 01 jesse-prod

Remembers facts you state and prose corrections you send with your key; they stay with that key. Code corrections are kept for review only: coding answers come only from Jesse's checked programs.

M / 02 jesse-pristine

The fixed baseline. Nothing you send changes it, so the same request returns the same answer. Use it for tests and comparisons.

D / 03 Memory

See what Jesse remembers for your key with GET /api/v1/memory; erase it with DELETE /api/v1/memory.

GET /api/v1/agent counts the pristine model's priors (this version and the three before it) and what your own agent has learned on top, or says it is fresh. POST /api/v1/agent/reset puts your agent back to the latest pristine copy; add {"documents": true} to remove your documents too.

D / 04 Documents

Upload up to 10,000,000 characters per document with POST /api/v1/documents and query it in later requests.

Open beta: be careful with confidential or export-controlled material. Your data is partitioned by API key and is not used to train other customers' results.

The numbers.

Storage limited, for now

Pristine model 7,259,598 Priors

Version b76e18b49217. Knowledge, grammar, research beliefs, web pages, Go and Wikipedia.

Largest customer model 7,259,648 Priors

The pristine model plus 50 priors one customer's agent has learned.

Request size, each 10M Characters

Characters per request and per document, about 2.5M tokens.

Average message latency 1 ms

Milliseconds, averaged over the last 1,000 chat messages, measured on the server.

MemoryIn progressGoal: 50 billion characters. Now: 500 MB of documents per account, searched on every request.
ContextIn progressGoal: no limit. Now: 10M characters per request; past that, your stored documents are searched.

Modalities

LanguageLiveChat and the API.
VisionIn progressNow: PDF pages and drawings in the vision lab. Coming soon: images in chat.
Audio (duplex)In progressNow: two-way room audio with speech recognition on our servers. Coming soon: Jesse's own voice, made on the server.
Voice (duplex)LiveTalk with Jesse, alone or with other people and up to four Jesses. Jesse's voice comes from your browser.
RoboticsLive, simulatedBodies, arms and drones that learn to move.
GamingLiveGo and other games.
FormalizationIn progressNow: research claims are checked in Lean 4 before they are published. Coming soon: formal proofs in chat.
Panel of ExpertsComing soonNot built yet.

One plan.

Developer / monthly

$20

Billed monthly by Stripe

Subscribe

KEYSUp to five API keys per account. Create, name and revoke them yourself.
RATEOne request every 200 ms per key. No daily or monthly request cap.
SIZE10,000,000 characters of context per request and per document. Past 100,000 characters Jesse reads the beginning and end in full and searches the whole text for the sections relevant to the question; the reply says so.
MODELSjesse-prod and jesse-pristine.
BILLINGChange your card, download invoices or cancel from the Stripe portal on your account page.

Endpoints.

Base URL https://jesse.my/api/v1. Authenticate with Authorization: Bearer jesse_live_….

Method	Path	Purpose
POST	`/chat/completions`	OpenAI-compatible chat completions. Supports `stream` and `tools`.
GET	`/models`	List the available models.
POST	`/feedback`	Mark an answer right or wrong (`rating`: `negative`/`thumbs_down` or `positive`/`thumbs_up`), with an optional correction. The response says what happened to it: `applied_to` and `effect`. A prose correction on `jesse-prod` is remembered for your key; a code correction is kept for review and does not change coding answers.
GET	`/memory`	Facts Jesse remembers for your key: things you state in chat requests ("my project uses Postgres", "remember that …"), plus your uploaded documents.
DELETE	`/memory`	Erase everything Jesse remembers for your key.
POST	`/documents`	Store a document for retrieval in later requests.
GET	`/documents`	List your stored documents.
POST	`/documents/query`	Search your stored documents.

Errors use the OpenAI shape. 401: missing or unknown key. 403: key revoked or subscription inactive. 429: slow down; the Retry-After header says how long.