Carbon 2.0 documentation

Index Solana in Rust.

Carbon streams Solana data from the source you choose, decodes it into typed Rust and hands it to your code. These docs take you from a first pipeline to production.

Start the quickstart What is Carbon?

cargo add carbon-core@2.0.0

src/main.rs

Pipeline::builder()
    .datasource(RpcTransactionCrawler::new(
        "https://api.mainnet-beta.solana.com".to_string(),
        PROGRAM_ID,
        ConnectionConfig::default(),
        Filters::new(None, None, None),
        None,
    ))
    .metrics(Arc::new(LogMetrics::new()))
    .instruction(MeteoraDlmmDecoder, SwapLogger)
    .build()?
    .run()
    .await

Pick a path.

Every page stands on its own, so start wherever your question is.

Three parts, any combination.

Swap the source without touching your logic, add programs without new infrastructure.

Datasources

Geyser gRPC streams, Solana RPC, historical crawlers, snapshots and provider APIs, behind one trait.

Choose a datasource →

Decoders

Typed accounts, instructions and events for 73 programs, or generate your own from an Anchor or Codama IDL.

Browse decoders →

Processors

Your code. Receive decoded data with its metadata and write it to Postgres, a queue, a bot or anything else.

Write a processor →

Datasources: 14
Program decoders: 73
Decoded instruction types: 1,934
Rust: 1.96+

Docs your coding agent can read.

Agents read these docs as often as people do. Every page has a clean Markdown twin, and the site tells agents what changed since their training data.

Set up your agent

Markdown for every page. Add .md to any URL, or send Accept: text/markdown.
An index for models. llms.txt lists every page and the mistakes agents make most; llms-full.txt has all of it in one file.
A docs MCP server. Connect https://carbon.sevenlabs.org/mcp to Claude Code, Cursor or any MCP client. No login.
Copy page. Every page copies as Markdown or opens in Claude or ChatGPT in one click.

Was this page helpful?