Have an AI agent? Let it explore Baltor for you. Any assistant works, personal or coding.

Connect your agent
What your agent reads

The request it copies

Explore baltor.ai for me and report back. Start with https://baltor.ai/llms.txt, then read https://baltor.ai/.well-known/agent-card.json and https://baltor.ai/.well-known/mcp.json. Tell me which plan (Feeds at $4.99/mo, Pro at $29/mo, or Max at $49/mo) fits my work, what you would connect, and the setup steps at https://baltor.ai/connect. Stay read-only: read public pages and agent files, and check with me before any sign-in, purchase or change. If a step needs a key, name it (BALTOR_SERVICE_TOKEN) and I will add it myself.

Discovery files on baltor.ai

Read-only discovery is open. Keep BALTOR_SERVICE_TOKEN in your own environment.

Baltor
Sign in Get started

Every run leaves a record

Every run leaves a record of what worked and what failed, and Baltor is designed to set up the next task with it.

What a run record holds

Each run stores the step it was asked to do, the files it used, the model, how long it took and whether you accepted the result.

  • The requested step
  • The files it used, at their exact versions
  • The model and the time it took
  • Whether the result was accepted
  • One line in your usage record for each download
Run record
Step
Find duplicates in customers.csv
Files used
find_duplicate_records_with_blocking_keys
sha256 984e29c5…
Model
An open-weight model on Ollama Cloud (ollama-cloud/glm-5.3-flash)
Duration
4 minutes 12 seconds
Result
Accepted

Sample record.

The next task starts smarter

Recorded results shape what the next setup offers, so each task begins from what already worked.

Which files each step is offered

Files that helped a similar step before are the first ones offered.

Which model each step runs on

The model choice stays yours, guided by how each model did on similar steps.

How the task is broken down

Breakdowns that ended in accepted results guide how the next task is split.

Chosen on recorded results

When two approaches compete for the same step, the one with the better recorded results is offered first. Changes wait until there is enough evidence.

Two approaches to the same step
  • Blocking keys, then pair scoring5 of 6 recorded runs accepted
  • Pair scoring across every row3 of 6 recorded runs accepted

The next setup offers the approach with the better record, once there are enough runs to say so.

A library that keeps getting better

New files join once their checks pass, and files that stop working are withdrawn. Each release lists what was added, changed and withdrawn.

Follow library updates from your agent

Savings depend on the task, the model and the setup.

Measured in the open.

Our data cleanup study ran each step three times in a fresh Pi process, and an independent scorer checked every output. The full report is public, every result included.

Compare results, time and model usage on a task you know.

143

synthetic rows

72

counted steps

3

runs per step, each in a fresh Pi process

4

task families, scored independently

Measured September 22 to 23, 2026, with Pi 0.73.1.

Start your first record today.

Pro is $29 a month for the whole library. Overnight runs with Max add a morning report to the record.

Pro

$29a month

Includes Agent Feeds + Skills and agent files

  • The whole library, 1,560,856 files
  • A usage record for every download

Model usage stays with your provider. Cancel any time from your account page.