SpanInsight
SpanInsight • Autonomous Intelligence

Your data deserves effortless intelligence, not complexity.

Analyze datasets with Google Colab cloud compute or local Python. Ask questions in natural language, build bot-protected survey forms by voice, watch multi-step AI reasoning unfold, and generate live interactive web reports or Jupyter Notebooks instantly.

google_colab_kernel.ipynb
Ask a question or type Python code...
Executing in Colab...
Completed Surveys 1,842 ↑ 12% this week
Satisfaction Rate 94.8% ↑ 4.2% overall
Response Trends
0
Analyses Run
0%
Local Privacy
0
App Downloads
4.9
User Rating
Capabilities

Enterprise Data Science on Your Device

Seamless cloud Python compute, transparent AI reasoning, and instant publication tools.

01

Google Colab Cloud Execution Engine

Run complex pandas, NumPy, scikit-learn, and PyTorch workflows seamlessly on remote Google Colab GPU/TPU kernels. Get real-time ANSI colored terminal output, fast data previews, and zero device overheating.

02

Autopilot & Transparent Reasoning

Autonomous multi-step exploratory analysis at a single tap. Watch step-by-step AI thought accordions reveal intermediate reasoning before code runs, complete with self-healing error retries and executive takeaways.

03

Smart Voice Surveys & Bot Defense

Dictate or describe survey goals to create smart forms in seconds. Collect responses securely with built-in anti-bot human verification and import verified submissions into your Python DataFrame with one click.

04

Live Web Reports & Jupyter Export

Export complete analysis sessions to standard Jupyter Notebooks (.ipynb) for full reproducibility, or publish interactive visual reports to reports.spaninsight.com with one toggle.

05

Verified Answers & Findings Memory

Every AI answer runs on real data first, then a verifier checks it against the actual output — stamped Verified with exact figures. Confirmed insights become your project's Findings Memory.

06

Smart Forms, Live Editing & Data Briefs

Edit published surveys without losing a single response (smart merge), and compile verified analysis into one-tap Data Briefs you can share anywhere.

Workflow

Instant. Autonomous. Verifiable.

From raw data to executive clarity in three effortless steps.

01

Connect & Inspect

Connect to a Google Colab session or local Python runtime. Drop in CSVs or Excel sheets for instant schema snapshots, distributions, and null detection.

02

Analyze Autonomously

Ask questions in plain English, use voice dictation, or run Autopilot. Create survey forms by voice and import verified responses into your dataset. AI writes and runs real Python code while showing its full reasoning process.

03

Export & Share

Export full Jupyter Notebooks (.ipynb), download clean PDF/Word documents, or share interactive reports on the web with permanent links.

Zero raw data retention

Your datasets stay inside your own Google Colab instance or local device filesystem. SpanInsight operates with strict zero-retention transit: only anonymized statistical prompts travel to AI models, never your sensitive raw database files.

Direct Colab Isolation
Zero Data Retention on Servers
Anti-Bot Form Protection
Full Python Code Transparency
FAQ

Frequently asked questions

What is SpanInsight?

SpanInsight is an autonomous data science workspace. It connects to Google Colab cloud runtimes or local Python, letting you import datasets, ask questions in plain English, and generate production-ready charts, statistical models, voice surveys, and interactive web reports.

How does Google Colab integration work?

SpanInsight connects directly to Google Colab runtimes via secure OAuth authentication. Heavy computations, deep learning models, and complex datasets execute on Google's cloud infrastructure with zero battery drain on your phone or computer.

Can I export my work to Jupyter Notebooks?

Yes! You can export your entire notebook session with all code cells, narrative explanations, and execution outputs directly to standard .ipynb files to open in VS Code, JupyterLab, or Google Colab.

How does the credit system work?

AI-powered code generation and Autopilot features consume daily credits. You receive 50 complimentary credits daily, refreshed automatically every 24 hours. Manual Python code execution via the built-in editor is always 100% free and consumes zero credits.

Are survey submissions protected from bots?

Yes. All public survey forms hosted on f.spaninsight.com are protected by non-intrusive anti-bot human verification to ensure only legitimate human responses are captured.

Is my data private?

Your datasets are stored on your device, and all processing runs inside your own Google Colab account under your login — not on shared servers. Nothing is retained by SpanInsight after your session.

What platforms are supported?

Android via Google Play and direct APK download from this page, plus Windows and Linux desktop builds. Widescreen/tablet layouts included.

Can I use SpanInsight without AI?

Yes. Expert Mode gives you a full Python editor against your dataset with zero AI involvement and zero credit cost — AI is optional, your control is absolute.

Download

Get SpanInsight - Free

Includes 50 complimentary AI credits renewed every 24 hours. No registration required.

Official Release

Install App

Experience private, AI-powered data intelligence on your phone, tablet, Chromebook, or desktop.

Other builds: ARMv7 | x86_64 (Emulator)
Changelog

Release history

v2.0.0 - Google Colab Cloud Execution, Autopilot & Live Web Reports (August 2026)

The Verified Intelligence release. Insight mode becomes an agent that plans, executes, and verifies every answer against your actual data — plus Findings Memory, Data Briefs, live form editing, and a major speed overhaul.

Verified Answers That Show Their Work

  • Agentic Insight Mode: Ask a question and the agent plans a multi-step approach, executes each step on your cloud kernel, then a separate verifier confirms the answer against the executed outputs before anything is shown.
  • Verification Badge & Honest Gaps: Confirmed answers carry exact key figures; partial answers list precisely what could not be determined.
  • Findings Memory: Verified insights are remembered per project — Home shows "What we know", and every future prompt inherits past findings as established facts.

Share & Collect Faster

  • Data Briefs: One tap compiles verified analysis steps into a shareable report, with Save-as-HTML for printing or PDF sharing.
  • Live Form Editing (Smart Merge): Edit published surveys in place — kept questions preserve every collected response, removed questions' answers purge permanently, share links never change.
  • Analyze Responses as Real Datasets: Survey submissions ship to Analysis as a proper CSV dataset, auto-refreshing when new responses arrive.
  • Quick Import: Paste a public CSV URL (fetched directly on the VM) or raw CSV/JSON straight from the create button.

Speed & Reliability

  • Persistent Kernel Connection: One live session per run instead of per-cell connect/teardown — first response and every step come back faster.
  • Big Excel Support: A high-performance reader auto-installs per VM — 500k+ row workbooks load in seconds, with parquet snapshots for instant reloads.
  • Offline Safety: First-run offline shows a retry screen instead of trapping you; resume pings the session so dead connections surface immediately.
  • Desktop Builds: Windows and Linux builds now ship alongside Android.
v1.2.3 - Local Dataset Caching & UX Improvements (June 2026)

Datasets now cache locally on import and reload silently on app restart. Stale caches auto-purge after 7 days to keep storage lean.

v1.2.2 - Expert Code Terminal & Smart Credit Flows (June 2026)

Introduces the Expert Code Terminal: a full VS Code-style expandable code editor for direct Python execution at zero credit cost.

v1.2.1 - Report Editor Layout (June 2026)

Cleaner Report Editor action flow with a prominent "View Live Report" button, standardized refresh styling, and an empty-Reports navigation fix.

v1.2.0 - UI/UX Refactoring & Layout Enhancements (June 2026)

Forms dashboard alignment, onboarding safe areas, standardized rounded action buttons across every screen, Android-optimized CI, and locally packaged fonts for faster offline launches.

v1.1.0 - Security & Performance Release (May 2026)

Sandbox builtins restriction, seed-phrase masking, parameterized SQL, 95% AI prompt compression, 10x faster data preview, adaptive glassmorphism, and a native boot screen.

v1.0.0 - Initial Release (May 2026)

Your personal data intelligence workspace: import data, ask questions, get answers — no code required.