Every word of
every document.
Instantly.
Every document your business touches — bid packages, specifications, plans, addenda, reports, assessments — contains critical information buried across hundreds of pages. Finding it today means opening files manually, scanning line by line, and hoping you remember where you read it. Document Search ends that.
You send us your documents. We process, index, and embed every character using local AI infrastructure — no third-party cloud, no data leaving a controlled environment. The result is a private, searchable intelligence layer built on your exact corpus. Every word of every document becomes instantly retrievable in under a second.
Mission Bay High School
Modernization Project.
This is not a demo environment.
A landscape and irrigation contractor is running Document Search on an active public works project — right now — using it daily to search plans, specifications, addenda, and compliance documents. The system is live, searchable, and in production.
Two modes. One corpus. Every answer.
Most search stops at exact words. Document Search goes further — into the meaning of what you wrote, the intent behind the question, the clause you're describing even when you don't know its name.
Every character of every page indexed using FTS5 BM25 — the same algorithm powering Elasticsearch. Type a term and see every mention across every document instantly, with surrounding context, page number, document type, and relevance score. Results return in milliseconds regardless of corpus size.
Ask questions in plain English against the meaning of the documents — not just the words. The system surfaces the relevant clause even if your exact words never appear in the document. Ask about concepts, requirements, obligations — the way you actually think about them.
Search results open into a live working environment.
Every result is a door. Click it and the document opens in a three-tab workspace — View, Full Text, Notes — alongside your search results. No new tabs. No switching applications. Everything in one terminal.
The original document opens at the exact page your search result came from. No manual hunting. No scrolling through 400 pages to find the clause. The system jumps directly to the evidence.
Every character extracted and rendered as searchable, selectable text. Built-in Ctrl+F finds any term instantly. In-document semantic search surfaces meaning within the single file. Ask a question about this document specifically.
Notes written against a document stay attached to that document permanently. Every observation, every flag, every question — searchable in the Notes Ledger alongside every document in the corpus. Your intelligence compounds with every session.
Every document passes through a quality verification layer. Clean text renders fully. Partially extracted documents display an Unverified Extraction notice so you always know the confidence level of what you are reading. Nothing ambiguous is presented as verified.
45 minutes of research. 90 seconds of output.
The Context Pod collects passages from multiple documents in a single session. When you have gathered what you need — clauses from six different specs, compliance requirements from three addenda, scope items from a bid package — one click copies everything to your clipboard. Ready for Claude, ChatGPT, or any AI for instant analysis, summarization, or email drafting.
Search any term or ask any question. Expand a result to read the surrounding document text with full context. Keyword highlighted, page-precise, sourced to the original document.
Click + Add to Pod on any passage that matters. Collect up to 15 passages from across as many documents as you need. The Pod holds everything — document title, source, and full text — with attribution intact.
Click Copy All. Everything in the Pod — sourced, attributed, formatted — is in your clipboard. Paste into Claude or any AI. Ask it to summarize the scope, identify conflicts, draft the email, compare the requirements. What took 45 minutes now takes 90 seconds.
Your team is paying to search for things they already have.
Industry research places information search friction at 30–40% of a knowledge worker's day. Run the math on your own payroll.
Conservative figure. Time spent opening documents one by one, searching file names, hunting for the clause, the number, the date. Information the business already owns.
Six full work weeks. Paid at full labor cost. Recovered nothing. No output. Pure friction inside a system with no index.
At $35/hr burdened labor. A 5-person office loses $42,000 per year to document search friction. The owner pays for every minute of it.
Eight capabilities. One deployed system.
Not a roadmap. Not a beta. A working pipeline in daily production use.
Every character indexed using FTS5 BM25 — the same algorithm powering Elasticsearch. Sub-second search regardless of corpus size. Type a term. See every mention across every document instantly.
Every result tagged with the exact page number the match came from. Click the result — the document opens at that page. No manual hunting. No Ctrl+F through a 400-page plan set.
PDF native and scanned, Word, HTML, plain text, email, images. Scanned pages run through a four-method OCR cascade. Every document type flows to the same searchable index.
478 canonical trade terms across six scope categories automatically identified in every document with page numbers and context. The system understands your industry, not just your keywords.
Collect passages across multiple documents and export them into Claude or ChatGPT with full source attribution. 45 minutes of research becomes 90 seconds of output.
Bid numbers, due dates, job walk dates, county, license requirements, prevailing wage flags, liquidated damages, bond percentages, engineer's estimate — all extracted automatically.
The system runs on your hardware. Documents never leave your network. No cloud dependency. One Python command and the server is live on your office LAN — every employee accesses it from any browser.
Bid due dates and job walk dates extracted automatically from documents appear in a calendar view. Per-job notes — timestamped, authored, searchable — permanently attached to every job record.
At a price you won't flinch at.
No per-user seat fees. No IT overhead. No implementation surprises. One setup fee, one monthly rate. Your documents — searchable in under a second.
- Full-text semantic search
- Page-precise results with PDF viewer
- Multi-format ingestion pipeline
- Per-job notes and audit trail
- Domain scope detection
- Everything in Starter
- BidDocument structured extraction
- Calendar intelligence — auto bid dates
- Context Pod — Claude/ChatGPT export
- Per-document notes with full-text search
- Prevailing wage, PLA, CWA auto-detection
- Everything in Standard
- Dedicated infrastructure
- Specialty Requests
- Embedded Graph Intelligence
- Custom domain enrichment models
Your documents are a library with no index.
Document Search builds the index.
Standard Terminal was founded on one conviction: the standard of information access in this country is too low, and price is the primary reason. Every estimator, every attorney, every office manager drowning in documents they already own deserves to find what they need in under a second. That is not a luxury. That is a standard. And we intend to make it universal.
Standard Terminal connects your documents, entities, and supplier records into a single machine-readable layer. Document Search clients receive priority access to Ledger Audit and Entity Profile. Reach out to learn what that unlocks →