Pairon
About Pairon and our guides
What Pairon does, who it is for, how our workflow guides use sources and examples, and where to find product evidence, limitations and corrections.
Updated
Pairon is a multiplayer AI coding workspace for people using Claude Code, Codex and OpenCode. A founder can bring a collaborator into a task, and a small engineering team can contribute to a shared agent conversation, keep decisions visible, and review the result together.
A shared conversation around local agents
Agents run on computers your team connects, using the agent account on the computer that runs the instruction. The computer owner chooses allowed folders and permissions. Pairon provides the shared conversation, project context and coordination around that work.
Start with two people and one task. Add parallel lanes when the work has independent parts and a clear review owner. The product walkthrough explains conversations, lanes and handoffs; the agent overview explains account use and supported agents.
The Free plan includes two people and one project. Agent-vendor subscriptions are separate from Pairon. Read current plan limits and billing details before choosing a plan.
How our guides are written
“By Pairon” identifies product documentation published by Pairon. These guides describe our own product and explain workflows around it; they are not independent reviews of Pairon or its competitors.
Each guide links to the product details and vendor documentation behind its claims. A link to a vendor describes that vendor’s behavior; it does not mean the vendor endorses Pairon. Agent features and account requirements can change, so check the linked documentation when setting up a workflow.
Examples show how to state a task, contribute a constraint, or hand off unfinished work. They are labeled illustrative. Room replays and the shared demonstration image use fictional people and example data; they are not customer stories or evidence of a measured productivity improvement.
Publication and modification dates describe the guide’s content history. A substantive change updates its modification date. A guide date alone is not a claim that every supported agent version, operating system or team arrangement has been tested.
What our evidence supports
The Codex project-memory benchmark reports its tasks, measurement method, uncertainty and limitations. It measures a particular agent setup; it does not establish a universal saving, a Claude Code or OpenCode result, or a measured improvement for a team.
The homepage’s team calculator is a model. Its assumptions are separate from the benchmark. Use the published methods when evaluating a numerical claim, and measure your own workflow before treating a demonstration as an expected result.
Our workflow guides and handoff template are available without an account. The template formats your notes locally. It does not inspect a repository, run tests, or reconstruct a lost agent session.
Understand the data boundary
Pairon stores conversation text, instructions and agent replies. File contents, diffs and terminal output are relayed live rather than stored as those payloads. Conversation messages and agent replies can themselves contain code or sensitive information. Local agents also communicate with their vendors according to their own settings and terms.
Read computer permissions and data handling and the privacy notice before connecting a project. Folder restrictions, approvals and redaction have limits; review what you send and what an agent proposes.
Questions and corrections
For product questions or a correction to a guide, email contact@pairon.dev. Include the public page URL, the statement you want us to check, and a source or reproduction when available. Do not send credentials or private repository contents.
Operator and service details are in the terms. A Pairon byline does not assert an outside certification or an independent endorsement.