We asked the LandingAI team what they think about Agentic Document Extraction (ADE) Second Generation. Document grounding came up most. ADE Second Generation reads a document hierarchically, from a section to a paragraph to a line to a single value. That granularity is what makes citations precise. One bounding box per rendered line lets you trace any value back to its exact spot on the page. For regulated teams, that changes the review loop. One healthcare customer runs millions of documents a month, with 70 to 80 people verifying results by hand. Cost came up next. Customers now pick a service tier, and DPT-3 Verity brings high accuracy to simpler documents without the full agentic cost. The team also pointed to the work customers never see. Evals, checkbox accuracy, and the distance between a prototype and a production stack. Full clips from the team in the comments.

ADE Gen2 offers a precise location for every line, every word and every table cell. We call it atomic grounding. When every word is localized, you can build some amazing things that were not possible before. ✅ Redact PII and PHI with exact localization ✅ Diff document versions to catch what changed ✅ Create UIs for rapid human-in-the-loop verification Read the full announcement on the blog (link in comments). Create an account and get started for free at ade.landing.ai

The documents you process every day just got 25% to 80% cheaper. That drop comes from three changes to pricing in Agentic Document Extraction (ADE) Gen2. With this release we are introducing: Complexity-based pricing: You only pay for the content a page returns, so a page with minimal content costs the minimum and a dense one scales up Service tiers named `standard` and `priority`: Pages you need now cost the most, pages that can wait cost the least. Standard tier is the default and costs 50% of Priority. Parse model choice between DPT 3 Pro and DPT-3 Verity (in public preview): The Verity model is tuned for digital text and tables at a fraction of the Pro model price. Cost is highly transparent. Each API response shows the total cost and the cost inputs. Read the full announcement on the blog. Try it on your own documents at ade.landing.ai

Today, we are announcing the second generation of Agentic Document Extraction (ADE). It is faster, more accurate, and more affordable than ever before. ADE Gen2 delivers higher performance while optimizing cost for every document. Pricing has been completely overhauled. Customers running mixed workloads should see 25% to 80% cost reductions. Groundings and citations now reach the word level following a page > block> line > word hierarchy. This allows every extracted value to point to the exact location on the page it came from. The outputs from the v2 Parse and Extract APIs are fully agent-ready and agent-friendly. Your agents get a clear hierarchy, stable IDs, and clean, standardized Markdown designed just for them. ADE Gen2 is a step change in document intelligence not an incremental update. Read the full announcement on the blog (link in comments). Create an account and get started for free at ade.landing.ai

Most document extractors hand your code untyped JSON and wish it luck. The values arrive as a plain object your code has never seen typed. A field goes missing, or its shape shifts, and nothing complains. The break surfaces at runtime, deep in the pipeline, far from the parse call. Agentic Document Extraction (ADE) ships an official TypeScript library that closes this gap. You define the fields once as a Zod schema. Parse returns the document as Markdown. Extract pulls those fields back typed to your schema. TypeScript infers the shape from there. A wrong field name or a missing value surfaces in your editor, before the code runs. The library carries zero runtime dependencies. The same typed pipeline runs anywhere JavaScript runs, from Node to browsers to the edge. Large documents get an async jobs path with a built in wait helper. Repo and quickstart in the comments.

Give your agent four ways to use ADE correctly from the first call. Ask a coding agent to call an API it was never taught and it improvises, inventing endpoints and skipping best practices. These four resources make sure your agent uses Agentic Document Extraction (ADE) right instead. The ADE CLI runs Parse and Extract from your terminal, no code. Built-in caching hashes your environment, parameters, and content, so identical re-runs pull cached files and skip duplicate API charges. The ADE Claude Skill teaches Claude how to call ADE inside your workflow. Two commands to install, then plain English prompts for tasks like parsing a batch of PDFs into structured CSV fields. The documentation MCP server opens a live connection to ADE's docs, so your agent pulls the latest reference on demand. It serves the docs, not the ADE tools. And llms.txt gives an LLM a lightweight ADE reference it loads in a single shot. Four resources, one goal. Your agents using ADE right from day one. Full walkthrough on YouTube, link in the comments.

Loosening the guardrails on AI spend made sense before anyone knew the cost. That is roughly where Dan Maloney, CEO of LandingAI, lands in this clip from the Data Science Dojo podcast with Raja Iqbal. His read is that companies pulled back the controls early, because the only way to learn what the tech could do was to let it run. Then the bills arrived. Spin up a thousand agents, keep them working around the clock, and the spend climbs fast. His line on it is simple. Spend five billion and change the game, and no CFO complains. Short clip below. Full episode in the comments.

A parser can read every word of a contract and still miss whether it was signed. Traditional OCR treats a stamp or a signature as noise, or as an unlabeled image blob. So a system can transcribe a permit or a certificate of origin end to end and never register whether it was signed, stamped, or sealed. For teams processing regulated paperwork, that is the one answer that matters. Agentic Document Extraction (ADE) closes the gap with attestation detection. At the parse stage, it finds these marks, assigns each a type, transcribes the text inside, and records where it sits on the page. Every mark comes back as one of four labels: SIGNED, E-SIGNED, STAMPED, or SEALED. Labels stack when a mark carries more than one, so an official filing stamped and signed in the same spot returns as [STAMPED][SIGNED]. The result is a queryable field, with grounding coordinates down to the line. "Is this document signed?" becomes a value your code can read, which lets you gate straight-through processing and route missing marks to a human automatically. One note for builders: attestation detection runs on DPT-3 Pro. Full breakdown, plus two real document examples, in the comments.