Training apps
Turn any screenshot into a drill. Show the solver line next to the one the player chose.
Early access · PLO4 and PLO5
Rangefinder is an API that reads a poker table from a screenshot (cards, board, stacks, pot, action) and returns the solver's play with frequencies and EV. One HTTP call, structured JSON back.
100 free requests with your key. No payment details.
200 OK
Sample request and response for illustration. The numbers are not from a live solve.
How it works
POST a PNG, JPEG or WebP of the table. A full-table screenshot is enough. No hand history, no parser, no plugin on the player's machine.
Rangefinder finds Hero's cards, the board, positions, stacks, pot and bets, and sends back exactly what it read.
Every available action with its frequency and EV, from a pre-solved library when the spot is covered and from a live solve when it is not.
The response
A wrong card read gives a confident wrong answer. So every response carries the read next to the solution, and says how close the solved spot is to the real one.
readread.confidencesolution.matchexact when the tree, stacks and board are the same. approximate when we mapped to the nearest solved spot, with a note on what changed.error{
"id": "slv_7f3k2m9q",
"read": {
"game": "plo5",
"street": "turn",
"hero": {
"position": "BTN",
"cards": ["Ah", "Kd", "Qh", "Tc", "7s"],
"stack_bb": 102.25
},
"board": ["Js", "9h", "4d", "2h"],
"pot_bb": 21,
"to_call_bb": 0,
"players": 6,
"in_hand": 2,
"line": { "preflop": "F-F-R-C-F-F", "flop": "B-C", "turn": "X" },
"confidence": 0.97
},
"solution": {
"match": "approximate",
"note": "Solved at 100 bb stacks. This hand started at 112 bb.",
"source": "library",
"actions": [
{ "action": "check", "frequency": 0.46, "ev_bb": 11.86 },
{ "action": "bet", "size_pct_pot": 50, "frequency": 0.17, "ev_bb": 11.88 },
{ "action": "bet", "size_pct_pot": 100, "frequency": 0.37, "ev_bb": 11.89 }
]
},
"timing_ms": { "read": 210, "solve": 38, "total": 262 }
}
curl https://api.rangefinderapi.com/v1/solve \
-H "Authorization: Bearer $RANGEFINDER_KEY" \
-F image=@table.png
import os
import requests
with open("table.png", "rb") as image:
response = requests.post(
"https://api.rangefinderapi.com/v1/solve",
headers={"Authorization": "Bearer " + os.environ["RANGEFINDER_KEY"]},
files={"image": image},
timeout=10,
)
spot = response.json()
for a in spot["solution"]["actions"]:
print(a["action"], a.get("size_pct_pot", ""), a["frequency"], a["ev_bb"])
import { readFile } from "node:fs/promises";
const form = new FormData();
form.append("image", new Blob([await readFile("table.png")]), "table.png");
const response = await fetch("https://api.rangefinderapi.com/v1/solve", {
method: "POST",
headers: { Authorization: "Bearer " + process.env.RANGEFINDER_KEY },
body: form,
});
const { read, solution } = await response.json();
console.log(read.board, solution.actions);
Integration
Reading a table and tracking a hand is most of the work in any solver product. Rangefinder does both, so you ship the part your users see.
/v1/solve/state.hand_id and the action line is tracked for you.max_wait_ms limits how long a live solve may take before you get the best answer available.Built for
Turn any screenshot into a drill. Show the solver line next to the one the player chose.
Let users drop in screenshots from a session and get an answer for each decision, without importing hand histories.
Annotate a student's hands in seconds and share each spot with frequencies and EV attached.
Feed delayed broadcast frames to the API and put solver frequencies in the overlay or the commentary notes.
Batch-process frames into structured spots and build your own datasets and reports.
Tell us what you are building in the access form. It decides what we ship next.
Request accessCoverage
Games
PLO4PLO5 cash games.
Hold'emPLO6Tournaments are not built yet. Vote for them in the access form.
Tables
Heads-up to 6-max.
Streets
Preflop, flop, turn and river.
Lines
Common lines are answered from the exact tree. Rarer lines and unusual stack depths are mapped to the nearest solved spot and marked approximate.
Speed
Target: under one second end to end for spots in the library. A live solve takes a few seconds and is capped by max_wait_ms.
Table layouts
A starter set of common online layouts. Tell us which ones you need when you request access.
Pricing
Start with 100 free requests. After that, buy a pack and use it when you need it. No subscription, no seats.
Try every endpoint on your own tables.
$0
Top up after the trial and keep testing.
$15
For a product with its first users.
$99
For volume and batch work.
$699
Early-access pricing, which may change before general availability. Need more than 10,000 requests? Tell us in the access form. Details on the pricing page.
Questions
Rangefinder is in private early access. We onboard developers in small groups and send keys in batches. Request access and tell us what you are building.
Early access starts with a set of common online layouts and grows on request. Screenshots work best when the whole table is visible, the image is at least 900 pixels wide, and nothing covers the cards.
Every response includes the parsed table and a confidence score, so you can check the read against the image or drop low-confidence answers. An image that cannot be read returns an error, not a guess.
A library of pre-solved spots answers most requests. When a spot is not in the library we solve it live, and the response says which of the two you got.
Yes. /v1/solve/state takes the cards, board, stacks and action line as JSON and returns the same solution object.
Rangefinder is built for study, review, training and broadcast tools. Most poker operators prohibit real-time assistance during play, and you are responsible for following the rules of any platform you or your users play on. We may revoke keys that are used to break them. See acceptable use.
Early access
We are onboarding developers in small groups. Tell us what you are building and how much you expect to call the API. It decides who we onboard first and what we build next.
Questions? Reach out to @we_will_give_a_username on Telegram.
We will email you when your key is ready. In the meantime, the docs show what the API returns.
Read the docs