{"slug": "show-hn-nace-ai-document-intelligence-1-0", "title": "Show HN: Nace.ai Document Intelligence 1.0", "summary": "Nace.ai launched Document Intelligence 1.0 (NDI) alongside its Drex decision model, offering both through a single API key and the nace-sdk npm package. NDI parses 36 file types and cites the source page and bounding box for every extracted value, while Drex returns a calibrated probability on each option in one forward pass, with a 128k context window and open weights described as coming soon. In a terminal demo, the system split a 38-page AP packet into 5 documents and routed a $9,870.00 invoice to payments at 0.93 confidence under Net 45 terms set by Contract amendment No. 3.", "body_md": "# Execute your workflows with Nace Small Models\n\nOne API key for Drex, which puts a probability on every call, and NDI, which reads 36 file types and cites the source of every answer.\n\nWorks with\n\n`npm install nace-sdk`\n## Builders on Drex, in their words.\n\n$499,235 of unbilled work across 1000 work orders. @NaceAI's drex found $482,260. drex reads the note and the invoice and flags what's missing. code does the math. when drex isn't sure, a human decides.\nDrex: faster, cheaper, better than jev, 32k context and with open weights coming soon? I will surely give it a try!\nI’m using Drex as a fast decision layer inside my project Elyreon/NOVA mainly for routing tasks, choosing tools/models, classifying failures, retries, urgency, and handoffs. the bot already responds sub 1s so i cant imagine how much better drex will make it\nNace shows Drex playing a strategy game. i gave it a harder opponent: my timeline Spoiler Guard sends each post to the new decision model and blurs the spoilers, 352 post checked in one session with sub-second calls\nUsing a massive LLM to decide which tool to call feels like hiring a CEO to sort your inbox. Drex targets those decisions: routing, ranking, classification and guardrails.\n\nSwitching was easy. I pointed my existing Jev setup at Drex, kept the same SDK, changed 3 environment variables, and it worked. I see a good fit for agent routing, tool selection, document classification, and choosing the next action in browser automation.\nsomeone wired a small open model to a real security camera 389 ms per call the vision model sees, drex decides\n128k context means it can route with real state instead of tiny summaries. the interesting part isn’t more generation. it’s better decisions before generation starts.\n\n## See it work from a terminal.\n\nAn agent with a Console key reads an AP packet, finds the page that matters and decides what to do with it. Every call in the window is a Console endpoint, and every value comes back with the page it was read from.\n\n$ export NACE_API_KEY=\"nace_sk_…\"\n\n$ claude\n\n> This AP packet from Halden Supply mixes invoices, a contract amendment and a W-9. Which invoice can we pay today, and did the terms change?\n\nconsole · split(file: \"halden-packet.pdf\")\n\n38 pages → 5 documents\n\n1 Invoice INV-2291 p. 1–3\n\n2 Invoice INV-2304 p. 4–6\n\n3 Contract amendment No. 3 p. 7–19\n\n4 W-9 p. 20\n\n5 Purchase order PO-77120 p. 21–38\n\nconsole · classify(document: 3, classes: \"procurement\")\n\ncontract_amendment · 0.97\n\nconsole · extract(document: 3, schema: { payment_terms, effective_date, parties })\n\npayment_terms \"Net 45, 2% discount within 10 days\" found\n\neffective_date \"2026-10-01\" found\n\nparties Halden Supply Co., Meridian Logistics found\n\nconsole · ground(document: 3, text: \"Net 45\")\n\npage 11 · box 0.12 0.41 0.88 0.47 · exact · 0.99\n\nconsole · extract(documents: [1, 2], schema: { invoice_no, total, payment_terms, po_number })\n\nINV-2291 $18,420.00 \"Net 30\" PO-77120 p. 2\n\nINV-2304 $9,870.00 \"Net 45\" PO-77120 p. 5\n\nconsole · systemone(state: \"INV-2304 · $9,870.00 · Net 45 · matches PO-77120 · terms per amendment No. 3\", questions: { approve: noul, route: choice })\n\napprove noul 0.93\n\nroute payments 0.88 · review 0.09 · reject 0.03\n\nPay INV-2304 today: $9,870.00 against PO-77120 on Net 45, the terms Contract amendment No. 3 sets on page 11, effective 1 October 2026. Drex puts 93% on approving it and routes it to payments.\n\nHold INV-2291. It still carries Net 30 on page 2, which predates the amendment. Want me to flag it for AP review?\n\n## Everything your agent needs, behind one key.\n\nAuthorization: Bearer $NACE_API_KEY\n\nOne key, one credit and one error envelope for both series.\n\n- ### DecisionDrex puts a calibrated probability on every option, in one forward pass. [Explore Drex](https://www.nace.ai/drex)\n- ### PerceptionNace Document Intelligence reads any file and points every value back to its source. [Explore NDI](https://www.nace.ai/ndi) POST /v1/documents/{task}invoice_2023T210.pdf\n  1. Parse # Invoice 2023T210 Item Amount Bookkeeping 600.00 Total 600.00 Markdown and layout blocks\n  2. Split \n    - Invoicep. 1–3\n    - Amendmentp. 4–16\n    - W-9p. 17\n Page ranges per document\n  3. Classify \n    - invoice0.98\n    - receipt\n    - bank_statement\n    - w9\n A label from your taxonomy\n  4. Extract \n    - /invoice_number2023T210found\n    - /total600.00found\n    - /po_numbernullnot found\n Fields, found or not found\n  5. Ground Total 600.00p. 1 The page and box it came from\n 36 file types: PDF, scans, Word, PowerPoint, Excel, CSV, email, HTML, images, audio, video Read by OCR, transcription and NER models.\n\n## Works where your agent works.\n\nConsole is plain HTTPS and JSON. Any harness that can call a tool can call it; these are the ones we test with.\n\n[CodexCodex calls the same endpoints from its sandbox; the key travels as one environment variable.](https://console.nace.ai/docs/developer-tools/mcp#codex)\n\n[CursorWire Console into Cursor's agent so a document question is answered inside the editor.](https://console.nace.ai/docs/developer-tools/mcp#cursor)\n\n[Claude CodeGive Claude Code a Console key and it splits, extracts and decides from the terminal.](https://console.nace.ai/docs/developer-tools/mcp#claude-code-and-claude-desktop)\n\n[opencodeopencode reaches Console over plain HTTPS; no plugin to install, nothing to keep in sync.](https://console.nace.ai/docs/developer-tools/mcp#opencode)\n\n## Three ways in.\n\n[REST APIWhenYour agent or service already speaks HTTP.You getPOST /v1/systemone for decisions and POST /v1/documents/{task} for files. Bearer key, JSON in, JSON out, one error envelope, idempotency keys on every job.](https://console.nace.ai/docs/drex/quickstart)\n\n[SDK, CLI and MCPWhenYou build in TypeScript or Python, work from the shell, or give a coding agent the tools.You getOne client for Drex and every document tool, with typed questions, uploads, job polling and retries built in. Code written for the TypeSafe SDK runs once you rename the imports.](https://console.nace.ai/docs/drex/sdks/python)\n\n[Your own deploymentWhenYour data cannot leave your network.You getThe same models and the same API inside your VPC, or air-gapped. Your keys, your logs, your IAM; updates pull in, nothing flows out.](#deployment)\n\n## One platform. Four boundaries.\n\nSame models, same API, same serving stack at every tier. The only thing that changes is the boundary around your data, from our cloud to your metal.\n\nYour data trains your models. The weights, adapters and eval sets stay yours, inside the boundary you pick.\n\n  - Where it runs\n  - Our cloud\n  - Network path\n  - Public TLS\n  - GPU tenancy\n  - Shared\n  - Updates\n  - Continuous\n ### Managed CloudStart on our GPUs, behind our API, with a key from Console. \n  - Shared GPU pool with strict tenant isolation\n  - Public API over TLS, no training on your data\n  - Usage-based pricing, no infrastructure to run\n For: pilots, evaluations, mid-market teams moving fast.\n  - Where it runs\n  - Our cloud\n  - Network path\n  - Private link\n  - GPU tenancy\n  - Dedicated\n  - Updates\n  - Continuous\n ### Dedicated CloudYour own GPU cluster in our cloud, reached over private links instead of the public internet. \n  - Single-tenant compute, pinned to your region\n  - PrivateLink or VPC peering into your network\n  - Customer-managed keys, dedicated throughput\n For: regulated teams that are cloud-first but done with shared infrastructure.\n  - Where it runs\n  - Your cloud account\n  - Network path\n  - Inside your VPC\n  - GPU tenancy\n  - Dedicated\n  - Updates\n  - Pull from registry\n ### Your CloudThe platform installs inside your AWS or Azure account. Data stays within your boundary. \n  - Runs in your VPC with your keys, your logs, your IAM\n  - Updates pull from our registry; nothing flows back out\n  - The deployment model our audit and banking customers run in production\n For: banks, Big 4 firms, anyone whose security review starts with “show us the network diagram.” [Talk to sales](https://www.nace.ai/contact?product=cloud&source=developers)\n  - Where it runs\n  - Your datacenter\n  - Network path\n  - None\n  - GPU tenancy\n  - Dedicated\n  - Updates\n  - Signed offline bundles\n ### On-Prem & Air-GappedYour datacenter, your hardware, zero outbound. \n  - No egress, no telemetry, offline license activation\n  - Signed offline update bundles on your schedule\n  - Ships as software on your servers or as a pre-configured appliance\n For: government, defense, and sovereign environments where the network cable isn’t there. [Talk to sales](https://www.nace.ai/contact?product=cloud&source=developers)\n\n## Questions\n\n## First request today. Your own cloud when you need it.\n\nCreate a key in Console and send your first request in a minute. When the data cannot leave your network, we bring the platform inside it.", "url": "https://wpnews.pro/news/show-hn-nace-ai-document-intelligence-1-0", "canonical_source": "https://www.nace.ai/developers", "published_at": "2026-10-09 00:58:25+00:00", "updated_at": "2026-10-09 01:16:57.719010+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-products", "ai-tools", "ai-agents", "large-language-models"], "entities": ["Nace.ai", "Drex", "Nace Document Intelligence", "nace-sdk", "Halden Supply Co.", "Meridian Logistics", "Jev"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/show-hn-nace-ai-document-intelligence-1-0", "markdown": "https://wpnews.pro/news/show-hn-nace-ai-document-intelligence-1-0.md", "text": "https://wpnews.pro/news/show-hn-nace-ai-document-intelligence-1-0.txt", "jsonld": "https://wpnews.pro/news/show-hn-nace-ai-document-intelligence-1-0.jsonld"}}