Baltor Get started

Harness and agent optimized operation

The right setup for every step your agent takes.

Baltor gives the AI harness you already use source-backed research and ready files for each step. Search the library, check a file's source and licence, and download the exact version. Keep your own model access.

Setup guides cover Claude Code, Codex, OpenCode, Pi and the Baltor Harness. Cursor, VS Code and other MCP apps connect to the same address.

An example working folder for one step
working-folder/
├── AGENTS.md
├── .agents/skills/split-address-lines/
│   ├── SKILL.md
│   └── scripts/split_address_lines.py
└── .mcp.json

Each file is checked against its digest

The problem

Is your AI spending effort in the wrong places?

Coding harnesses and agent systems lose effort in the same ways, whatever the task.

  • Too much context for a small task

    Steps carry the whole history, so a small step gets slow and confused.

  • An expensive model for every decision

    Simple transformations go to the largest model because nothing picks a smaller one.

  • Missing domain expertise

    The agent guesses at methods a specialist already knows.

  • Rewriting code that already exists

    Tested code gets written again, and paid for again as output.

  • The same mistakes, again

    A correction made in one run does not reach the next one.

  • Long work small models cannot finish alone

    Large, multi-step problems stall without the right files at each step.

Source: the customer problems in AGENTS.md and README.md.

What Baltor is

Reusable files for technical and creative work

Choose files for the task: reusable implementations, reference data, prompts, workflows and checks. Each item keeps its supporting files.

  • Instruction files

    AGENTS.md, CLAUDE.md and other context files

  • Skills

    SKILL.md with its scripts, references and assets

  • Tools and reusable code

    Tested scripts and code a step can run

  • Data and workflows

    Schemas, reference tables and native workflow files

  • Prompts and task briefs

    Reusable instructions, examples and expected results

  • MCP server settings and plugins

    The tools a step may call

Your models, your keys

Use your own model provider or local endpoint. Library access and model costs are separate.

Checked before it is served

Every item starts as a candidate. Automated checks of licence, source, safety and declared effects run before a file is published. Execution evidence varies by item.

A fresh harness for each step

The Baltor Harness already runs a task on your model in small steps. Starting a fresh standard harness for each step is not open yet.

Source: README.md, what Baltor gives your harness, one harness for each step, and what is not open yet; the harness file definition in AGENTS.md.

How it works

Search returns short results. A step downloads only what it chose.

  1. Search

    The harness asks the library for what a step needs. The answer is a short list of results: name, purpose, source, licence, exact version and checksum. No file contents.

  2. Choose

    The step picks one result. An item that needs file, network or process access is offered only to a step that declares that access.

  3. Download and check

    Access is checked again at download. The file must be exactly the version the step chose, and the download is recorded in your usage.

  4. Place

    The file goes where the harness reads it: a skill folder, AGENTS.md or a protocol server entry.

Source: README.md, searching and retrieving and usage and what you pay for.

One step getting one file from the Baltor library The harness searches the library and gets back short results only. The step chooses one and downloads it. The library checks access again and returns the exact bytes. The harness checks the bytes against the digest and places the file in the step's working folder. Your harness at one step Baltor library harness components search the library results only name · licence · version · digest chooses one download that item access checked again the exact bytes recorded in your usage exact chosen version placed where the harness reads it: .agents/skills/split-address-lines/SKILL.md

Available today

A working library and an editable creative demo

The library delivers files. Your chosen harness does the work. The demo shows a separate creative project you can explore.

1,560,856 files

Distinct files in the library, from 256,380 harness components. Identical files count once.

Source: the live count the service reports when this page opens; before that, the October 9 publication record

Search and complete downloads

Inspect the source, licence, version and required permissions. Download the selected item with its supporting files.

Source: the service component guide

Your existing harness

Connection guides for Claude Code, Codex, OpenCode, Pi and the Baltor Harness; Cursor, VS Code and other MCP apps use the same address. Automatic placement and manual setup have different support.

Source: the connection guide records

Keep the project editable

The demo export was imported, rendered and reopened in Blender. That proves this project path, not automatic video recreation.

Source: the creative proof and its limits

Account and usage

Sign up, connect a scoped client key and inspect metered usage. A complete download activity timeline is not available yet.

Source: current behavior and next steps

Measured in the open

We test our own library, and publish the result when it says no.

A data cleanup study compared library items with no item, on a design frozen before the first model call and scored by an independent checker. Each step ran in a fresh, isolated harness process that held only its own files.

225 of 400

Requests used of the declared ceiling, across 72 counted steps. Every request reported its token usage.

Source: REPORT-2026-09-22.md

0.842 against 0.991

An approved phone number item made a cheap model clearly worse: 0.842 in all three repetitions, against a mean of 0.991 without it. The item's method was incomplete, and the report names the rows it lost.

Source: REPORT-2026-09-22.md

0 of 4

Task families where the library items were clearly better; they were clearly worse in 1. The items cost 45 to 446 percent more prompt tokens per step.

Source: REPORT-2026-09-22.md

Next: a harder population, where the model fails without help, reusing the same frozen design, scorer and isolation recipe. The report asks for the measured harm to go into the item's review record; a revised item needs a new review and a new measurement.

The library plan

The served library and its growth targets

Recorded supply and future targets are separate. Files count only when an active item in the library delivers them; repeated files count once.

  1. Observed

    1,560,856

    distinct files in the library today

    Source: the live count, or else the October 9 publication record

  2. Observed

    256,380

    harness components in the same library, each with a distinct reusable job

    Source: the live count, or else the October 9 publication record

  3. Plan

    Two million

    distinct served files, the next milestone after the first million

    Source: the file-supply roadmap

  4. Aim

    One million

    useful harness components, the longer-term supply target

    Source: the current north star

The next proofs

  • Customer benefit. Harder data, debugging, infrastructure and creative tasks, with and without Baltor. Keep unsuccessful results.

  • Reference video. The next proof: explain a permitted clip, find the tools and prepare an editable variation. No automatic recreation today.

  • Public Good. Free components for sustainable development and public-benefit work, available with an account.

  • Faster, traceable retrieval. Measure wrong files, stale sources and latency before choosing a cache or a Rust engine.

Source: the ordered delivery and business plan.

Business model

Library access, with your own model provider

Agent Feeds gives agents source-backed research. Agent Feeds + Harness Files pays for access to the hosted library, and the local overnight tools come with it. You supply your model access and choose where the work runs.

  • The first 10 accounts from Baltor's sign-up hold Agent Feeds + Harness Files free each month.

    Source: the release 24 record and account_policy.py

  • Agent Feeds is $9 a month, free through December 31, 2026 (Eastern), with no automatic charge. A paid feed subscription needs the customer's opt-in.

    Source: the Agent Feeds pricing decision

  • Live payments through Stripe checkout and the customer portal since September 21, 2026.

    Source: the payments decision in AGENTS.md

  • Public Good access is available. An enabled account can download the selected free files within fair-use limits. Bring your own model access and execution environment.

    Source: the Public Good access policy

Agent Feeds + Harness Files

$49 a month

  • The whole library, searchable from your harness
  • One item at the version you chose counts once a month, with no overage billing
  • A Baltor key for each tool you connect
  • Cancel any time from your account page
Get started

Source: README.md, the current plan and billing unit

Why now

The formats became standards. Choosing and trusting the right file did not.

33.9 to 50.5 percent

Curated skills raised the mean pass rate from 33.9 to 50.5 percent on 87 tasks in SkillsBench, June 2026. Skills an agent wrote for itself scored 8.1 to 11.5 percentage points below no skills at all.

Sources: SkillsBench on arXiv, as recorded in the first-party package research

157 of 98,380

Skills confirmed malicious in a February 2026 study of 98,380 skills from two registries. Supply is plentiful; a reason to trust a file is not.

Sources: the study on arXiv, as recorded in the harness and library plan

Baltor's answer: one library, delivered one step at a time, with each file's licence and exact version checked before it is used.

Contact

Talk to us.

Try the playable demo, explore the library, or create an account and follow the setup guide for your harness. For help, use the support page.

Baltor.AI 1428 Bryn Mawr St Saxton, PA 16678 United States Source: the privacy notice, PRIVACY-NOTICE.md