PARALLEL DISPATCH.
Inference jobs are fanned out across live network nodes simultaneously. The coordination layer routes each request to the best available GPU tier and returns results in parallel.
A terminal coding agent. Inference dispatched across decentralized GPU nodes, verified with a SHA-256 Proof-of-Execution, paid in $PRLX credits. It runs open-source models, lower quality than frontier services.
Dispatches inference in parallel, attaches a verifiable commitment to every result, and settles in $PRLX credits with no per-call gas.
Inference jobs are fanned out across live network nodes simultaneously. The coordination layer routes each request to the best available GPU tier and returns results in parallel.
Every call returns a SHA-256 commitment over the request, result, and node identity. You know which node served it. A hash commitment, not a zero-knowledge proof.
Authenticate with a pk_live_ key from the Console Developers page. The endpoint is OpenAI-compatible. $PRLX credits are metered off-chain: deposit once, no per-call gas.
Install the CLI, authenticate with your key from the Console Developers page, then open the interactive session.
parallelix-code auth pk_live_<your-key>Paste your key from ai.parallelix.io/developers. Stored locally, never sent anywhere except the network endpoint.
parallelix-codeOpens the interactive session. The agent connects to the network and waits for your task.
refactor the auth middlewareType a task. The agent reads your files, dispatches inference across GPU nodes, writes the edits, runs tests, and prints the PoE line.
Three stages from your terminal to a verified result on the network.
You type a task in your terminal. The agent reads the relevant files in your working directory.
The coordination layer dispatches inference across live GPU nodes in parallel. Multiple nodes can handle sub-requests simultaneously.
A Proof-of-Execution comes back with every result: a SHA-256 commitment over the request, result, and node identity. Verifiable on-chain.
Available now. Get a key at the Console Developers page, then run the CLI.