ModelsParticula-Finance
Figures from filings,down to the footnote.
Particula-Finance extracts figures from filings, statements, and decks, including the tables and footnotes where the real numbers hide. Every value is typed and carries a source span. Self-hosted, so material non-public information stays yours.
YOUR-GPU · PARTICULA-FINANCE
$ POST /v1/extract · 10-K_2025.pdf
→ particula-finance · figures locked
→ "revenue": "$84.2M" · "yoy_growth": "+18%"
→ "ebitda": "$12.6M"
✓ 200 · 48 ms · 99%+ field extraction
ready
PARTICULA-FINANCE · YOUR HARDWARE · ZERO TELEMETRY
What it doesSee it on your data
Figures from filingsRevenue, margins, segment data, and guidance from 10-Ks, statements, and investor decks.
Tables and footnotesLayout-aware parsing for the places disclosures actually live: multi-level tables, footnotes, and restated comparatives.
A source span on every numberEach value links back to page and line in the source document, so analysts audit the extraction instead of redoing it.
Numbers stay numbersTyped output with currency, units, and periods normalized. No string soup: figures land in your model ready to compute.
Spec sheetCompare all six
Context window32K tokensA full annual statement section fits in one request.
Median latency48 ms medianBatch 1 on the reference single-GPU deployment. Our own measurement: the shipped harness reproduces it on your hardware.
InterfaceOpenAI-compatible RESTServed by vLLM, deployed with Docker Compose or Helm. Drop-in for existing OpenAI client code.
Hardware1× 24 GB GPUL4 or A10 class. Sizing reviewed before the pilot.
Fine-tuningLoRA on your dataTuning jobs are included with Enterprise. A few hundred labeled examples from your pipeline is usually enough to move the needle.
How deployment worksStart at step 01
01Discovery call
30 minutes, no slides. Your task, your data shape, and the security constraints around it. If this model is the wrong fit, we say so on the call.02Pilot install
Containerized deployment to staging in your VPC or on-prem, on the same stack as Lumen and Notetaker. The eval harness runs on your data and the numbers go into the pilot agreement.03Production rollout
Runbook handoff, operator training, and 90 days on call. After that it is yours to operate, no vendor dependency.More modelsAll six models
Request a demo
“Bring a hundred real examples from your pipeline. We run the model on them live, the harness scores the output in front of you, and you decide with numbers instead of a sales deck.”
Sebastian Mondragon, Founder of Particula Tech
Request a demo30 minutes · no sales deck