TypeScript

Use Evolve’s TypeScript client to run Codex on one DeepSWE task and save the results. The SDK submits the job to Evolve and follows it from your program.

Install and authenticate

You need Node.js and npm and a Full access key from the dashboard’s API keys page. In your project directory, install the SDK and tsx to run the example. Replace your-api-key with your key.

npm install @evolvingmachines/evolve
npm install --save-dev tsx
export EVOLVE_API_KEY="your-api-key"

See installation for account setup.

Run one task

Save this as eval.mts. It selects only anko-default-function-arguments from deepswe@1.1 and runs one trial with Codex and GPT-6 Luna.

eval.mts

import { hosted } from "@evolvingmachines/evolve";

const evolve = hosted();

const job = await evolve.jobs.start({
  datasets: [{
    name: "deepswe",
    version: "1.1",
    task_names: ["anko-default-function-arguments"],
  }],
  agents: [{
    name: "codex",
    model_name: "gpt-6-luna",
  }],
  max_trial_spend_usd: 1,
  retry: { max_retries: 0 },
});
console.log("Job:", job.id);

const finished = await evolve.jobs.watch(job.id);
console.log("Status:", finished.status);
console.log("Model cost:", finished.stats.cost_usd);

for await (const trial of evolve.jobs.trials(job.id)) {
  console.log(
    trial.task_name, trial.status, trial.reward
  );
}

const archive = await evolve.jobs.download(job.id, {
  to: "./results",
});
console.log("Saved:", archive);

Run it in the shell where you set the API key:

npx tsx eval.mts

The program prints the job ID, final status, model cost, trial result, and saved archive path under ./results/.

Call	Result
`jobs.start()`	Returns the accepted job before the trial finishes.
`jobs.watch()`	Waits for the final job status.
`for await (... of jobs.trials())`	Reads trials across all pages.
`jobs.download(..., { to: "./results" })`	Saves the job’s `.tar.gz` archive and returns its path.

A completed job can contain failed trials. Read the trial’s status and reward; a missing value is not zero.

The example sets a $1 model-spend cap for this trial attempt and disables configured infrastructure retries. A model call already in progress can finish above the cap. Provider capacity waits can still occur. Closing the program does not cancel the job.

Show progress

Replace the program’s watch() call with one of these snippets, using the same evolve client and job. Use an event callback while awaiting completion:

const finished = await evolve.jobs.watch(job.id, {
  onEvent: (event) => console.log(event.seq, event.type),
});

Or consume events directly:

for await (const event of evolve.jobs.watch(job.id)) {
  console.log(event.type, event.data);
}
const finished = await evolve.jobs.get(job.id);

Choose one consumption style for each watch handle.

Read a trial

Add this after the job finishes, using the same evolve client and job. It reads the trial’s normalized trace and its stored-file listing.

const page = await evolve.jobs.trials(job.id, {
  limit: 1,
});
const trial = page.items[0];

if (trial) {
  const trajectory = await evolve.trials.artifact(
    trial.id, "trace-atif"
  );
  const files = await evolve.trials.files(trial.id);
  console.log(trajectory, files.items);
}

Awaiting jobs.trials() reads one page; for await reads every page. The stored-file listing is also one page. Follow its nextCursor when present; see trial files.

HTTP API

Python and TypeScript

Run agents

Shared behavior

SDK workflows

SDK resources

SDK methods

Install and authenticate

Run one task

Show progress

Read a trial

Next step

Configure a job

Inspect files