OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

Carbon

Carbon is a Python course and codebase for building a coding-agent harness incrementally. Its 15 chapters introduce agent components, demonstrate the capabilities they enable, and show examples using a real model.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Carbon — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Carbon as a guided codebase when learning agent-harness engineering or exploring concrete implementations of individual harness primitives.

Who it is for

  • Engineers comfortable with Python who want to learn how coding agents work by building one.
  • Readers seeking harness-engineering concepts they can apply beyond a single agent library.

Documented uses

  • Learn how a model call becomes a tool-using coding agent.
  • Study how instructions, project context, history and context limits affect an agent workflow.
  • Explore patterns for sandboxed command execution, durable sessions, orchestration and subagents.
  • See how candidate work can be checked against assertions and how calls can be traced.
Before you choose

Requirements & limits

The repository is public and lists an MIT license.

  • The project is written in Python and expects familiarity with Python.
  • Model access is through an OpenAI-compatible endpoint; the README names LM Studio, OpenRouter and Ollama.
  • The chapters cover both agent capabilities and engineering concerns such as approval gates, execution environments, persistence and verification.
  • The repository includes an interactive command-line agent, chapter demos and a Textual terminal interface.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up Carbon with a model endpoint

Install the project dependencies and configure an OpenAI-compatible model endpoint.

Before you start

  • Python project environment managed with uv
  • An OpenAI-compatible endpoint; the README gives LM Studio, OpenRouter and Ollama as options

Steps

  1. Run the command below to create the environment and install dependencies.

    Documented command
    uv sync

  2. Copy .env.example to .env.

  3. Set LLM_BASE_URL, LLM_MODEL and LLM_API_KEY for the endpoint you intend to use.

Expected outcome

The repository is configured to connect through its model provider package to an OpenAI-compatible endpoint.

Guide limitations

  • These steps cover setup; they do not document endpoint installation or model selection.
  • The README says the optional reasoning-effort, provider-order and quantization variables apply only behind a multi-provider router.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Agent orchestration

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
CarbonLearn how a model call becomes a tool-using coding agent.CLIThe repository is public and lists an MIT license.
AgeniteDefine typed tools with JSON Schema validation and error handling.CLI
VeldraCreate an agent with a policy, tools, skills, knowledge base and reasoning method.Self-hostedThe documented deployment is self-hosted and uses a sign-in gate after a first-run administrator account is created.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
7 October 2026
Original website title
thecarbonlayer/carbon · Official homepage title · retained in its source language
How we review sourcesSuggest a correction