Basic run

To get an in-depth, cited answer to a question, start a run with the query, wait for it to finish, and read the result. This guide demonstrates that workflow in full. The quickstart shows the shortest path to a result. This guide walks the same request through in more detail, including what the SDKs’ create_and_wait / createAndWait hide from you, and the manual poll a plain REST caller has to do instead.

1. Start a run

create_and_wait / createAndWait call create, then poll until the run finishes, and hand back the completed envelope. cURL has no waiter, so the same request only starts the run:

from seltz import Seltz

client = Seltz(api_key="your-api-key")

run = client.agent.create_and_wait(
    "What are the three most notable AI-company acquisitions announced in 2026?"
)

import { Seltz } from "seltz";

const client = new Seltz({ apiKey: "your-api-key" });

let run = await client.agent.createAndWait({
  query: "What are the three most notable AI-company acquisitions announced in 2026?",
});

curl -X POST https://api.seltz.ai/v1/agent/runs \
  -H "x-api-key: $SELTZ_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "What are the three most notable AI-company acquisitions announced in 2026?"}'

The cURL call returns at once with the run in pending state. output is present but its fields stay unset until the run reaches a terminal state:

{
  "id": "run_01a072afeedf7df2bfe72350dc381785",
  "object": "agent.run",
  "status": "pending",
  "stop_reason": null,
  "created_at": "2026-09-05T17:48:47Z",
  "started_at": null,
  "completed_at": null,
  "request": {
    "query": "What are the three most notable AI-company acquisitions announced in 2026?",
    "output_schema": null
  },
  "output": {
    "text": null,
    "structured": null,
    "sources": [],
    "grounding": []
  }
}

2. Poll for the result

Retrieve the run by the id the create call returned. In Python and TypeScript, create_and_wait / createAndWait already did this polling for you above, so run there is already the finished envelope; get is the single check it repeats under the hood, useful on its own when you want to look a run up independently instead of holding onto the object create returned. Over cURL, this is the only way to find out whether the run is done:

run = client.agent.get(run.id)

run = await client.agent.get(run.id);

curl "https://api.seltz.ai/v1/agent/runs/run_01a072afeedf7df2bfe72350dc381785" \
  -H "x-api-key: $SELTZ_API_KEY"

See Poll for completion for the full loop that turns one check into “wait until done,” and for wait’s options. Check status before reading anything from output. A run that hasn’t finished yet comes back as running (or pending), not an error, and output’s fields stay unset, the same status as right after the run is created:

{
  "id": "run_01a072afeedf7df2bfe72350dc381785",
  "object": "agent.run",
  "status": "running",
  "stop_reason": null,
  "created_at": "2026-09-05T17:48:47Z",
  "started_at": "2026-09-05T17:48:47Z",
  "completed_at": null,
  "request": {
    "query": "What are the three most notable AI-company acquisitions announced in 2026?",
    "output_schema": null
  },
  "output": {
    "text": null,
    "structured": null,
    "sources": [],
    "grounding": []
  }
}

started_at distinguishes running from pending: it is null while a run waits for a worker, and set the moment one picks up the run. Keep polling until status is completed, failed, or cancelled, or use wait / createAndWait, which does exactly this loop for you. Once the status is completed, a completed run’s output.text is a cited Markdown report, with inline [n] markers pointing at output.sources:

{
  "id": "run_01a072afeedf7df2bfe72350dc381785",
  "object": "agent.run",
  "status": "completed",
  "stop_reason": "finished",
  "created_at": "2026-09-05T17:48:47Z",
  "started_at": "2026-09-05T17:48:47Z",
  "completed_at": "2026-09-05T17:49:05Z",
  "request": {
    "query": "What are the three most notable AI-company acquisitions announced in 2026?",
    "output_schema": null
  },
  "output": {
    "text": "## Three most notable AI-company acquisitions announced in 2026\n\nAssuming “notable” means a combination of deal size, strategic significance, and impact on the AI industry, the leading three are:\n\n1. **SpaceX → Anysphere (Cursor) — $60 billion** (June 16). SpaceX announced an all-stock agreement to acquire Anysphere, the company behind the AI coding platform Cursor, at an implied $60 billion equity value. The transaction would make Anysphere a wholly owned SpaceX subsidiary and represents SpaceX’s major expansion from aerospace and satellite communications into AI software. [1] Crunchbase described it as the largest startup acquisition on record and said it accounted for roughly half of 2026 U.S. venture-backed startup M&A spending at that point. [2]\n\n2. **Nvidia → Hugging Face — about $13 billion** (announced by September 3). Nvidia agreed to acquire Hugging Face, a major platform for AI models, datasets, and applications. It would be Nvidia’s largest outright acquisition, giving the chipmaker greater influence over AI-model distribution while expanding beyond chips into models, developers, and infrastructure. [3] Nvidia said Hugging Face would remain an open platform, with its more than 18 million developers retaining choice over models, cloud providers, and chips. [3]\n\n3. **Qualcomm → Modular — approximately $4 billion** (June). Qualcomm pursued/acquired Modular, an AI software and chip company whose platform lets developers deploy models across different chips without rewriting code. The deal is notable because it forms part of Qualcomm’s rapid effort to build an AI-infrastructure portfolio beyond mobile processors. [4] A contemporary M&A analysis ranked Qualcomm’s Modular purchase as the next-largest non-biotech U.S. startup acquisition after Cursor, at about $4 billion. [2]\n\n**Caveat:** The available sources are inconsistent on the exact status and terms of the Qualcomm–Modular transaction: one June 22 report said Qualcomm was still nearing a deal and that final terms had not been reached, while another later analysis described it as an acquisition. [4][2]",
    "structured": null,
    "sources": [
      { "id": 1, "url": "https://pulse2.com/spacex-to-acquire-cursor-in-stock-deal-valuing-ai-coding-startup-at-60-billion/" },
      { "id": 2, "url": "https://news.crunchbase.com/ma/2026-mergers-acquisitions-record-cursor-spcx/" },
      { "id": 3, "url": "https://www.hindustantimes.com/business/nvidia-to-acquire-hugging-face-for-13-billion-why-the-ai-model-deal-matters-101788440128258.html" },
      { "id": 4, "url": "https://ca.finance.yahoo.com/news/qualcomm-nears-4b-deal-ai-221210867.html" }
    ],
    "grounding": []
  }
}

3. Read the result

In Python and TypeScript, create_and_wait / createAndWait already returned this same envelope, so there is no separate call. Read run.output.text and walk run.output.sources for the citation list:

print(run.output.text)
for source in run.output.sources:
    print(f"[{source.id}] {source.url}")

console.log(run.output?.text);
for (const source of run.output?.sources ?? []) {
  console.log(`[${source.id}] ${source.url}`);
}

Example output

## Three most notable AI-company acquisitions announced in 2026

Assuming “notable” means a combination of deal size, strategic significance, and impact on the AI industry, the leading three are:

1. **SpaceX → Anysphere (Cursor) — $60 billion** (June 16). SpaceX announced an all-stock agreement to acquire Anysphere, the company behind the AI coding platform Cursor, at an implied $60 billion equity value. The transaction would make Anysphere a wholly owned SpaceX subsidiary and represents SpaceX’s major expansion from aerospace and satellite communications into AI software. [1] Crunchbase described it as the largest startup acquisition on record and said it accounted for roughly half of 2026 U.S. venture-backed startup M&A spending at that point. [2]

2. **Nvidia → Hugging Face — about $13 billion** (announced by September 3). Nvidia agreed to acquire Hugging Face, a major platform for AI models, datasets, and applications. It would be Nvidia’s largest outright acquisition, giving the chipmaker greater influence over AI-model distribution while expanding beyond chips into models, developers, and infrastructure. [3] Nvidia said Hugging Face would remain an open platform, with its more than 18 million developers retaining choice over models, cloud providers, and chips. [3]

3. **Qualcomm → Modular — approximately $4 billion** (June). Qualcomm pursued/acquired Modular, an AI software and chip company whose platform lets developers deploy models across different chips without rewriting code. The deal is notable because it forms part of Qualcomm’s rapid effort to build an AI-infrastructure portfolio beyond mobile processors. [4] A contemporary M&A analysis ranked Qualcomm’s Modular purchase as the next-largest non-biotech U.S. startup acquisition after Cursor, at about $4 billion. [2]

**Caveat:** The available sources are inconsistent on the exact status and terms of the Qualcomm–Modular transaction: one June 22 report said Qualcomm was still nearing a deal and that final terms had not been reached, while another later analysis described it as an acquisition. [4][2]

[1] https://pulse2.com/spacex-to-acquire-cursor-in-stock-deal-valuing-ai-coding-startup-at-60-billion/
[2] https://news.crunchbase.com/ma/2026-mergers-acquisitions-record-cursor-spcx/
[3] https://www.hindustantimes.com/business/nvidia-to-acquire-hugging-face-for-13-billion-why-the-ai-model-deal-matters-101788440128258.html
[4] https://ca.finance.yahoo.com/news/qualcomm-nears-4b-deal-ai-221210867.html

Next steps

Structured output: requesting a JSON result instead of Markdown
List runs: paging this org’s runs, newest first
Poll for completion: managing the wait yourself
Cancel a run: best-effort cancellation, and why it can lose the race to completion
Handle errors: rejected requests versus failed runs
Reference: endpoints, fields, limits and errors
Agent API Reference: full REST request and response specification

Getting Started

Web Tools

Data

Tutorials

Integrations

SDKs

1. Start a run

2. Poll for the result

3. Read the result

Next steps

​1. Start a run

​2. Poll for the result

​3. Read the result

​Next steps

1. Start a run

2. Poll for the result

3. Read the result

Next steps