Blacksmith AI

VULCAN Toolbox Suite

A folder of RFP PDFs in. A structured, compliant response out.

The Toolbox is a five-stage document workflow (Init, Digest, Analyze, Generate, Export) that turns a solicitation package into a working response. Every stage is explicit, inspectable, and billable: deterministic where structure matters, AI-driven where judgment matters, and always transparent about which is which.

Open beta · founding-customer pricing
vulcan.blacksmithai.net
Toolbox project creation with location, agency, and due date fields

Built For

Proposal teams staring at the PDF wall

A solicitation lands as a folder: the RFP, attachments, amendments, a PWS, wage determinations. The Toolbox ingests the whole package and builds the response structure the documents actually demand, before anyone writes a word.

Skeptics of black-box AI

If you can't see what the machine decided, you can't sign your name to it. Every Toolbox stage reports what it parsed, what it extracted, and what it generated, so review means checking work, not reverse-engineering it.

Teams shipping volumes at tempo

The Toolbox is how VULCAN ships proposal volumes on deadline: a repeatable pipeline where volumes generate in parallel, long jobs survive network hiccups, and progress is visible stage by stage instead of a spinner and a prayer.

Init: structure before a single word

Upload the solicitation package and the Toolbox initializes the engagement: a two-pass parse extracts the RFP into structured buckets, documents and volumes are built from the solicitation's actual architecture, and your org profile is normalized into a frozen snapshot every later stage reads from: one source of truth, established up front.

  • Two-pass RFP extraction sorts the package into structured requirement and instruction buckets
  • Volumes and sections built from the solicitation's real structure, not a generic template
  • A frozen snapshot normalizes org and solicitation text once, so every stage works from the same ground truth

Digest: every page, even the ugly ones

Government documents are hostile terrain: scanned PDFs, multi-column layouts, tables nested inside tables. The Digest stage parses PDF, DOCX, XLSX, PPTX, and images with OCR, with advanced document processing handling the complex layouts, so requirements buried on page 87 of a scanned attachment still make it into the analysis.

  • Parses PDF, DOCX, XLSX/CSV, PPTX, and images with OCR
  • Advanced layout handling for tables, multi-column pages, and scanned documents
  • Digested content flows into structured, searchable form instead of raw text soup

Analyze & Generate: judgment where it belongs

The Analyze stage decomposes requirements, maps them against your actual capabilities, and surfaces the gaps honestly. Generate then writes with volume-specialized prompts and section skeletons that carry page budgets and evaluation-criteria cross-walks, followed by quality agents that check voice consistency, page discipline, and compliance trace.

  • Requirements decomposition and compliance gap analysis against your real capabilities
  • Volume-specialized generation with per-section page budgets tied to evaluation criteria
  • Post-generation quality agents: voice consistency, page budget, and compliance trace checks

Export, plus a pipeline you can watch run

The final stage produces submission-grade files: DOCX with solicitation headers, confidential footers, and page numbers; PDF preview; Excel for structured deliverables. And because every long-running stage reports its phase in real time, you always know what the pipeline is doing, and jobs keep running server-side even if your connection drops.

  • DOCX export with solicitation headers, confidential footers, and page numbering; PDF preview; Excel output
  • Live phase reporting on every stage, never a frozen spinner hiding a dead job
  • Each stage is explicit and billable, so you can see exactly what ran, in what order, at what cost

VULCAN keeps our teams aligned, accelerates proposals, and brings clarity to every phase of execution… Together, they've become a trusted force multiplier across our government and commercial programs.

Joseph "JR" Rodriguez, President & CEO of CS3 EGS

Common questions

What actually happens at each of the five stages?

Init parses the solicitation package and builds the document structure it demands. Digest extracts clean, structured content from every file format in the folder. Analyze decomposes requirements and maps them against your capabilities. Generate drafts sections with page budgets and quality checks. Export produces formatted DOCX, PDF, and Excel deliverables.

Is the whole pipeline AI?

Deliberately not. Structure is handled deterministically (parsing, document assembly, formatting, export) because structure should be repeatable. AI is applied where judgment matters: requirements analysis, gap assessment, and drafting. The Toolbox is always transparent about which is which, which is exactly why the stages are explicit instead of fused into one magic button.

What file types can it digest?

PDF, DOCX, XLSX/CSV, PPTX, and images with OCR. Complex layouts such as scanned PDFs, multi-column pages, and dense tables go through advanced document processing rather than naive text extraction, which is where most RFP tooling quietly loses requirements.

What's shipped today versus on the roadmap?

Shipped: the full five-stage workflow producing RFI and Sources Sought responses and proposal volumes, with DOCX, PDF, and Excel export. On the roadmap: SBIR and white-paper output formats, plus standalone Strategize, Compete, and Analyze tools that can run independently or chain into the full capture workflow.

Run your next RFP folder through the Toolbox

7-day free trial. No upfront charge. Every feature included.