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

What Baltor is

Research, files and overnight tools for the agents you already run. You keep your harness, your models and your project.

Updated October 25, 2026

Three offerings, one account

Baltor gives the agents in your harness two things to work with, research for decisions and ready files for building, plus tools that keep long work going while you are away. Your harness is the program that runs your coding agent, such as Claude Code, Codex, OpenCode or Pi.

OfferingWhat your agent getsPlan
Agent FeedsSource collections for model, benchmark, tool and paper decisions, as JSON Feed, RSS and Markdown.Feeds plan, free through December 31, 2026 (Eastern), then $4.99 a month
Skills and agent files1,560,856 distinct files: skills, instruction files, tools and code, subagents, commands, hooks and MCP server settings.Pro, $29 a month, with Agent Feeds
Overnight workProcess supervisors for long runs on the Baltor Harness, with call budgets, time limits, checkpoints and a morning report.Max, $49 a month, Public preview

Each download names its source, licence and exact version, with a SHA-256 checksum. Identical files count once; the file count comes from the capabilities record, read October 25, 2026. Paid plans start when you opt in. The Public Good collection is free with an account, and the Baltor Harness is free and open source.

Compare plansBrowse the libraryExplore Agent Feeds

What stays with you

Baltor adds research and files to the setup you already trust. Three things stay exactly where they are.

Your model access

Your harness calls your own models with your own key. Your model keys stay with you.

Your agent

Your harness runs the work, with the permissions you choose. Baltor answers requests for material.

Your project

Your repository stays on your machine. You choose what to send, and the service answers what you ask.

Four kinds of useful material

Useful material covers how to approach the work, code that can do it, what happened before and what you want done differently. Each kind shows what it holds for one task: importing a customer file.

Context Intelligence

Methods, questions, constraints, examples and output contracts that guide the work.

A field-definition guide and a checklist for missing values.

Code Intelligence

Reusable functions, tools, files and workflows with declared inputs, outputs and effects.

A qualified normalizer with its dependency and verification records.

Runtime History and Solution Intelligence

Saved outcomes, failures, repairs, measurements and solution information that can inform a new task.

A previous import's failed assumptions and the checks that caught them.

User Feedback Intelligence

Scoped corrections, priorities and instructions supplied by a person.

Keep uncertain matches for review, and keep the original file as it is.

How a request flows

Every request runs through two short sequences, one on the service and one in your own tools.

On the service

  1. Check your access: find material your account may use.
  2. Find relevant help: search guidance, reusable code and previous findings.
  3. Deliver selected files: check that each download matches the version you chose.

In your tools

  1. Set the goal and limits: decide what success means and what the work may access.
  2. Carry out each step: use your chosen models, development tools or existing code.
  3. Keep the results: inspect files, checks and the record of what happened.

Searches return references; the body arrives with the download your agent chooses. Requests go through your signed-in connection, and your computer stays under your control. On the website, a unit of work is called a step, and each step can run in a fresh harness, so it starts clean and focused.

Ways to connect

MCP
https://baltor.ai/mcp, streamable HTTP, with OAuth or a Baltor key in the Authorization header. Tools: intelligence_search, provisioning_manifest, provisioning_read and public_good_files.
REST API
Versioned JSON routes under /api/v1/. Capabilities and health answer to anyone; the other routes take a Baltor key.
Agent Feeds files
/feeds/sources.json and /feeds/sources.md, one file per collection, and library updates as catalogue.json, .rss and .md.
Bash
# Ask the service what it serves
curl https://baltor.ai/api/v1/capabilities

Serving and connectionsSearching and retrieving

Licences and attribution

  • Each item names its own licence, source and exact version. Licences vary from item to item, so read the one on each item.
  • Keep the LICENSE and ATTRIBUTION.md files that come with an item.
  • Generated work carries the licence its record shows.
  • A downloaded file is material for your agent to read. An item that declares effects, such as writing files or running commands, says so before download.

Who runs Baltor

Baltor is operated by Baltor.AI, info@baltor.ai, United States. Contact is by post, at that address.

SupportTerms of servicePrivacy notice

Need a hand? Status, the connection checklist and the operator's email address are on one page.

Get support