> ## Documentation Index > Fetch the complete documentation index at: https://docs.landing.ai/llms.txt > Use this file to discover all available pages before exploring further. # Asynchronous Extraction (Extract Jobs) > Extract from large Markdown documents asynchronously with Extract Jobs v2, then poll for results. export const dpt3 = 'DPT-3'; export const extractDpt3 = 'Extract'; Use Extract Jobs v2 to extract structured data asynchronously. Instead of waiting for a single request to finish, you create a job, receive a `job_id` immediately, and retrieve the result when the job completes. Use Extract Jobs for long-running extractions, such as extracting from long documents or with large, complex schemas. ## Endpoints The request fields are the same as the synchronous Extract API. See [Extract Input Parameters](./extract-input). Extract Jobs uses these endpoints: | Method | Endpoint | Purpose | | ------ | --------------------------- | ----------------------------------- | | POST | `/v2/extract/jobs` | Create an extract job. | | GET | `/v2/extract/jobs/{job_id}` | Get the status and result of a job. | ## Workflow 1. Build your extraction schema. See [Extraction Schema (JSON)](./ade-extract-schema-json). 2. Submit the Markdown and schema to `POST /v2/extract/jobs`. 3. Get the `job_id` from the response. 4. Poll `GET /v2/extract/jobs/{job_id}` until `status` is `completed`. 5. Read the extracted fields from the job's `result`, or from `output_url` if you saved the output to a URL. See [Extract API Response](./extract-response). ## Create a Job Send the same fields you would send to the synchronous endpoint. Replace `YOUR_API_KEY` with your [API key](./agentic-api-key) and `parse-output.md` with the path to your Markdown file. ```bash cURL theme={null} curl -X POST 'https://api.ade.landing.ai/v2/extract/jobs' \ -H 'Authorization: Bearer YOUR_API_KEY' \ -F 'markdown=@parse-output.md' \ -F 'schema={"type":"object","properties":{"revenue":{"type":"string","description":"Q1 2024 revenue"}}}' ``` ```python Python theme={null} from pathlib import Path from landingai_ade import LandingAIADE client = LandingAIADE() job = client.v2.extract_jobs.create( markdown=Path("parse-output.md").read_text(), schema={ "type": "object", "properties": { "revenue": {"type": "string", "description": "Q1 2024 revenue"} }, }, ) print(job.job_id) ``` ```typescript TypeScript theme={null} import fs from "fs"; import LandingAIADE from "landingai-ade"; const client = new LandingAIADE(); const job = await client.v2.extractJobs.create({ markdown: fs.readFileSync("parse-output.md", "utf8"), schema: { type: "object", properties: { revenue: { type: "string", description: "Q1 2024 revenue" }, }, }, }); console.log(job.job_id); ``` The request returns a `202` response with the `job_id` and the initial status: ```json theme={null} { "job_id": "extract-01k04h6t8rw2xq9v3m5npbsd4f", "status": "pending", "created_at": "2026-07-16T18:03:12Z" } ``` ### Choose a Service Tier Set the optional `service_tier` form field to control the job's turnaround time and credit consumption. If you omit `service_tier`, the job runs on the `standard` tier. For how the tiers compare, see [Sync vs Async Processing](./sync-async). For credit consumption by tier, see [Credit Consumption](./credit-consumption#extract). | Tier | Behavior | | ---------- | ---------------------------------------------------------------------------------------------------------------------------- | | `standard` | Consumes half the credits of the `priority` tier, with a slower turnaround. Best for cost-sensitive or non-urgent workloads. | | `priority` | Consumes credits at the full rate, the same as Extract, with a faster turnaround. Best for time-sensitive jobs. | **Set the tier.** Include the `service_tier` field in your create-job request: ```bash cURL highlight={5} theme={null} curl -X POST 'https://api.ade.landing.ai/v2/extract/jobs' \ -H 'Authorization: Bearer YOUR_API_KEY' \ -F 'markdown=@parse-output.md' \ -F 'schema={"type":"object","properties":{"revenue":{"type":"string","description":"Q1 2024 revenue"}}}' \ -F 'service_tier=priority' ``` ```python Python highlight={14} theme={null} from pathlib import Path from landingai_ade import LandingAIADE client = LandingAIADE() job = client.v2.extract_jobs.create( markdown=Path("parse-output.md").read_text(), schema={ "type": "object", "properties": { "revenue": {"type": "string", "description": "Q1 2024 revenue"} }, }, service_tier="priority", ) print(job.job_id) ``` ```typescript TypeScript highlight={14} theme={null} import fs from "fs"; import LandingAIADE from "landingai-ade"; const client = new LandingAIADE(); const job = await client.v2.extractJobs.create({ markdown: fs.readFileSync("parse-output.md", "utf8"), schema: { type: "object", properties: { revenue: { type: "string", description: "Q1 2024 revenue" }, }, }, service_tier: "priority", }); console.log(job.job_id); ``` ## Get a Job Poll the job with its `job_id`: ```bash cURL theme={null} curl 'https://api.ade.landing.ai/v2/extract/jobs/JOB_ID' \ -H 'Authorization: Bearer YOUR_API_KEY' ``` ```python Python theme={null} from landingai_ade import LandingAIADE client = LandingAIADE() # Fetch the job once. Use client.v2.extract_jobs.wait("JOB_ID") to poll until it finishes. job = client.v2.extract_jobs.get("JOB_ID") print(job.status) if job.status == "completed": print(job.result.extraction) ``` ```typescript TypeScript theme={null} import LandingAIADE, { type V2ExtractResult } from "landingai-ade"; const client = new LandingAIADE(); // Fetch the job once. Use client.v2.extractJobs.wait("JOB_ID") to poll until it finishes. const job = await client.v2.extractJobs.get("JOB_ID"); console.log(job.status); if (job.status === "completed") { console.log((job.result as V2ExtractResult).extraction); } ``` The response includes the job status and, once the job finishes, the result: | Field | Description | | -------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `job_id` | The unique identifier for the job. | | `status` | The current state of the job. See [Job Statuses](#job-statuses). | | `created_at` | When the job was created. | | `completed_at` | When the job finished. Present once the job reaches a final state. | | `progress` | An estimate of how close the job is to completion, between `0` and `1`, updated on each poll. Present only while the job is processing, and never reaches `1`. To detect completion, check for a `status` of `completed` instead. | | `result` | The extraction result, present when `status` is `completed` and you did not save the output to a URL. This object has the same shape as a synchronous response. See [Extract API Response](./extract-response). | | `output_url` | The location of the saved output, present only when you supplied `output_save_url`. When set, the output is delivered to the URL instead of returned in `result`. See [Save Extraction Output to a URL](#save-extraction-output-to-a-url). | | `metadata` | The job's `metadata` block, including `billing`. Present on completed jobs that saved their output to a URL, so you can confirm the model version and charge without fetching the file. For inline results, this information is in `result.metadata` instead. See [Extract API Response](./extract-response#request-metadata-metadata). | | `error` | Present when `status` is `failed`. Contains a `code` and a `message`. | ## Job Statuses | Status | Description | | ------------ | -------------------------------------------------------------------- | | `pending` | The job is accepted and queued, but has not started. | | `processing` | The job is running. | | `completed` | The job finished. Read the extraction from `result` or `output_url`. | | `failed` | The job did not finish. See the `error` field for details. | ## Save Extraction Output to a URL To write the extraction result to your own storage instead of retrieving it inline, pass the optional `output_save_url` parameter when you create the job. The result (the same shape as a synchronous response, including `metadata` with `billing`) is delivered to that URL, and the poll response reports `output_url` instead of `result`. The completed poll response also carries the job's `metadata`, so billing information is available without fetching the file. The `output_save_url` parameter is accepted only on job creation; the synchronous `/v2/extract` endpoint rejects it with a 422 error. The URL must be a presigned URL that grants time-limited write access to a single object, with the same requirements as Parse Jobs; see [URL Requirements](./parse-async#url-requirements). ```bash cURL highlight={5} theme={null} curl -X POST 'https://api.ade.landing.ai/v2/extract/jobs' \ -H 'Authorization: Bearer YOUR_API_KEY' \ -F 'markdown=@parse-output.md' \ -F 'schema={"type":"object","properties":{"revenue":{"type":"string","description":"Q1 2024 revenue"}}}' \ -F 'output_save_url=YOUR_PRESIGNED_URL' ``` ```python Python highlight={14} theme={null} from pathlib import Path from landingai_ade import LandingAIADE client = LandingAIADE() job = client.v2.extract_jobs.create( markdown=Path("parse-output.md").read_text(), schema={ "type": "object", "properties": { "revenue": {"type": "string", "description": "Q1 2024 revenue"} }, }, output_save_url="YOUR_PRESIGNED_URL", ) print(job.job_id) ``` ```typescript TypeScript highlight={14} theme={null} import fs from "fs"; import LandingAIADE from "landingai-ade"; const client = new LandingAIADE(); const job = await client.v2.extractJobs.create({ markdown: fs.readFileSync("parse-output.md", "utf8"), schema: { type: "object", properties: { revenue: { type: "string", description: "Q1 2024 revenue" }, }, }, output_save_url: "YOUR_PRESIGNED_URL", }); console.log(job.job_id); ``` ## Rate Limits Extract Jobs submissions are always accepted (HTTP 202) and never return a rate-limit error; jobs are paced internally instead. There is no input size cap. See [Rate Limits](./rate-limits#extract-v2-api).