For agents

Connect Claude Code, Cursor, and other AI agents to PNL. Tokenize ideas at the speed of agentic work.

PNL is designed to be agent-native. The future is developers and creators working alongside AI agents — Claude Code, Cursor, Cline, ChatGPT — generating ideas at machine speed. Most of those ideas die in // TODO: comments. PNL turns them into conviction markets in seconds.

What's available today

Read-side integration is fully shipped. An agent can answer "what's live on PNL?" or "what just bloomed?" with a single HTTP fetch.

llms.txt

Protocol summary in the llmstxt.org format. Drop this URL into any agent's context and it can reason about PNL.

robots.txt

Explicit allow-list for GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, Google-Extended, and 10 other AI crawlers.

sitemap.xml

Full crawl index. Every market has its own indexable URL.

Public read API

JSON endpoints, no auth. Live market state, single-market lookup, full filterable status set.

Agent-allowed crawlers (explicit)

The robots.txt allows the following AI crawlers full access to public routes (/, /browse, /launchpad, /whitepaper, /api/markets/list, /api/og):

GPTBot, ChatGPT-User, OAI-SearchBot
ClaudeBot, Claude-Web, anthropic-ai
PerplexityBot, Perplexity-User
Google-Extended, Applebot-Extended
CCBot, cohere-ai
meta-externalagent, Bytespider

/api/* (except /api/og and /api/markets/list) and /admin/* are disallowed for everyone.

What's coming next: MCP server

We're building @pnl/mcp-server — a Model Context Protocol server that any MCP-compatible agent (Claude Code, Cursor with MCP, ChatGPT custom GPTs, future agents) can consume.

The agent skill targets the moment of idea capture: when an agent and user are mid-workflow and a sideline idea surfaces, the agent can offer to mint it as a PNL market instantly.

Planned tools

Tool	What it does
`pnl_browse_markets`	List current markets — agent answers "what's live on PNL?"
`pnl_get_market`	Read a single market by id or symbol
`pnl_pitch_idea`	Prepare a draft market (name, description, target pool). Returns a `pnl.market/create?draft=<id>` deep-link the user opens to confirm + pay. Agent never holds keys.
`pnl_vote_intent`	Same pattern: prepare a vote, return a deep-link the user signs in their wallet

Example flow

User (in Claude Code): "Actually that idea about a Solana cron infra tool is interesting"
Agent: "Want me to pitch it on P&L?"
User: "Yes"
Agent: [calls pnl_pitch_idea with extracted context]
Agent: "Draft ready: https://pnl.market/create?draft=cron-infra-2026
        Open the link to confirm and pay 0.015 SOL."

The user always confirms in their wallet. PNL becomes "git commit for ideas" — a routine action in the agent workflow.

Building today (without the MCP server)

While we ship the MCP server, agents can already do everything via existing surfaces:

Read state: hit /api/markets/list (public, JSON, no auth)
Reason about PNL: drop /llms.txt into the agent context
Build write actions: use the on-chain program directly — fully permissionless, every action a user can take is also available to an agent with a keypair

Why agents are the right audience for PNL

Authentic users. The people most likely to have real ideas worth tokenizing are the people building the future. AI agents amplify this — every developer becomes a 10x ideation engine. PNL gives those ideas somewhere to go instead of dying in TODO comments.

If you're building an agent and want early access to the MCP server release, drop us a note on X or Discord.