Does EdgeTal send candidate resumes or search data to cloud servers?

No. EdgeTal operates 100% on-device. All resume text extraction, 512d MediaPipe vector embedding generation, and Gemma-2B LLM fit reasoning execute locally inside your hardware sandbox without sending 1 byte of cloud telemetry.

How does EdgeTal run AI models without internet or token fees?

EdgeTal uses quantized Int4 Google Gemma-2B IT and MediaPipe Text Embedder bound directly to native Metal/GPU hardware delegates on iOS, macOS, and Android.

How can recruiters collaborate and share candidate pools privately?

Recruiters can export password-encrypted .edgetal packages and share them peer-to-peer over AirDrop, USB, or email.

iOS Public Beta is Live via Apple TestFlight!Connect your iCloud email to get instant access to build 1.0.1+6.

Join iOS Beta Guide →

Step-by-Step User Guide

How to Use EdgeTal

EdgeTal turns your smartphone, tablet, or laptop into a 100% self-contained AI recruiter. Follow this 5-step workflow to parse resumes, search candidates in sub-100ms, run explainable AI candidate fit evaluations, and collaborate with your team securely.

5-Step On-Device Recruiter Workflow100% OFFLINE CAPABLE

STEP 01

Import Resumes

PDF / DOCX / CSV

STEP 02

Create Job Role

Skills & Vector

STEP 03

Vector Search

Sub-100ms Matching

STEP 04

AI Fit Analysis

Gemma-2B Reasoning

STEP 05

Pipeline & Share

Encrypted .edgetal

Step 1: Import Candidates in Seconds

Multi-Source Ingestion Pipeline

Open the Import Screen in EdgeTal and select your candidate files using one of three ingestion modes:

From Folder: Select a local folder containing PDF or DOCX resume files on your mac, iPhone, or iPad.
Cloud Folder: Directly select files stored in iCloud Drive, Google Drive, or OneDrive.
CSV Import: Load structured candidate datasets via a local CSV file or direct CSV web URL.

⚡ Under the hood: EdgeTal extracts text, candidate names, emails, and automatically generates 512-dimensional vector embeddings on your device using MediaPipe Text Embedder.

Step 2: Create Open Job Roles & Competencies

Targeted Hiring Specs

Navigate to Jobs & Pipelines and tap + New Job. Enter the job title, target location, required skill tags (e.g. Flutter, Kafka, PostgreSQL), and paste the full job description.

EdgeTal instantly builds a 512-dimensional job vector representation in your local ObjectBox database, enabling instant candidate vector ranking against this specific role.

Step 3: Run Natural Language Talent Searches

Sub-100ms Vector Search

Instead of writing complex Boolean queries, type natural search phrases into the Search Bar (for example: “Senior Full-Stack Engineer with microservices, Flutter, and high-scale Postgres experience”).

EdgeTal performs sub-100ms dense vector matching, surfacing candidates ranked by semantic similarity score percentage while highlighting exact keyword occurrences.

Step 4: Execute On-Device Gemma-2B AI Fit Analysis

Explainable LLM Screening

Tap any candidate profile card, open AI Fit Analysis, select your target Job Role from the dropdown, and tap Run Fit Analysis.

Gemma-2B IT On-Device Analysis Outputs:

💭

Step-by-Step Thought Process: Clear reasoning showing how candidate experience maps to job requirements.

📑

Verifiable Evidence Extraction: Exact quotes and skill achievements extracted directly from the candidate CV.

📌

Verdict & Missing Skills: Definitive fit verdict (Shortlist, Maybe, Pass) with highlighted skill gaps.

Step 5: Manage Pipeline & Share Encrypted Packages

Peer-to-Peer Team Sharing

Drag candidates across your recruitment pipeline stages (Shortlisted → Interviewing → Offer → Placed).

When sharing candidate pools with recruiting team members, export a password-encrypted .edgetal archive from Backup & Export. Teammates can import the package over AirDrop, USB, or email without risking cloud data exposure.

Ready to test EdgeTal on iOS?

Join the Apple TestFlight Public Beta to test on-device talent intelligence.

Get TestFlight Access →