Skip to content

Latest commit

 

History

History
222 lines (152 loc) · 9.27 KB

File metadata and controls

222 lines (152 loc) · 9.27 KB

What is OpenAnalyst?

What's at the core of OpenAnalyst and how can you use it?

OpenAnalyst is an intelligent, AI-powered VS Code extension that transforms how data teams work. It merges AI-driven autonomous coding with reusable templates and specialized agents, enabling teams to collaborate, standardize workflows, and deliver insights faster than traditional methods.

Whether you're a data scientist, analyst, or business intelligence professional, OpenAnalyst helps you build shareable, transparent workflows that eliminate setup bottlenecks and accelerate time-to-insight—from weeks to hours.

Key Features

  • Specialized AI agents for every data role
  • Template-driven collaboration via shareable configurations
  • Natural language to code with 35+ intelligent tools
  • Instant team onboarding through auto-import settings
  • Inline visualizations that render directly in your workspace

Under the Hood

OpenAnalyst is powered by advanced LLMs (Claude, GPT-4, Gemini, and 15+ providers) with a unique multi-agent architecture. Each agent is a specialized expert—Data Analyst, Data Scientist, Data Engineer, Business Analyst, or Research Analyst—with domain-specific instructions and tool access.

It's designed to autonomously execute multi-step tasks, delegate to specialized agents when needed, and capture the entire process for team learning and iteration. OpenAnalyst also supports template-based knowledge sharing, making organizational standards instantly accessible to every team member.


Who Can Use OpenAnalyst?

Data Teams & Analytics Organizations: Who need to standardize workflows, share best practices, and onboard new members without weeks of training.

Startups and Lean Teams: Requiring powerful AI assistance for data work without building custom infrastructure or lengthy prompt engineering.

Cross-Functional Teams: (Product, marketing, operations) who want to collaborate on data-driven decisions with consistent methodologies.

Enterprise Data Departments: Looking to encode institutional knowledge, ensure compliance with data governance rules, and accelerate insight delivery across dozens of analysts.


OpenAnalyst Tools and Features

Templates in OpenAnalyst

Templates turn tribal knowledge into shareable assets.

OpenAnalyst Templates let you:

  • Define specialized agents with custom roles, instructions, and tool permissions
  • Share reusable prompts for common analyses (cohort analysis, churn prediction, A/B testing)
  • Encode organizational rules (PII handling, approved libraries, reporting formats)
  • Export and import configurations in one click via JSON
  • Auto-load team standards on startup with zero manual setup

Use Cases:

  • A data team lead creates a "fintech-analysis" template with 5 specialized agents and 20 reusable prompts
  • New team member joins, imports company-config.json, and is productive on day one
  • Marketing team shares a "campaign-analysis" template across regional offices for consistency

Templates make OpenAnalyst your team's shared knowledge base, accessible to every member with standards automatically enforced.


Agents in OpenAnalyst

Agents turn generic AI into domain experts.

OpenAnalyst Agents let you:

  • Deploy specialists for specific tasks (Data Analyst for exploration, Data Scientist for modeling, Data Engineer for pipelines)
  • Delegate complex workflows where an Orchestrator agent coordinates multiple specialists
  • Restrict tool access per role (e.g., Business Analysts can't modify production code, only documentation)
  • Follow proven methodologies encoded in each agent's custom instructions
  • Switch contexts instantly by changing agents mid-task

Use Cases:

  • A Data Analyst agent explores customer data and creates inline visualizations
  • When ML modeling is needed, it delegates to a Data Scientist agent via new_task
  • A Business Analyst agent calculates ROI while a Research Analyst gathers competitive intelligence
  • All results synthesized by the Orchestrator into a comprehensive report

Agents blend specialization, autonomy, and collaboration—making OpenAnalyst your AI data team.


Configuration Transfer in OpenAnalyst

Configuration Transfer turns hours of setup into seconds.

With Configuration Transfer, you can:

  • Export entire workspace settings including API keys, custom agents, and provider configs to JSON
  • Import configurations from teammates, merging settings without overwriting local customizations
  • Auto-import on startup by setting autoImportSettingsPath in VS Code settings
  • Version control configs in Git for team-wide updates and rollbacks
  • Share templates via YAML files containing agents, prompts, and rules

Use Cases:

  • A team lead exports their optimized setup: oa-code-settings.json
  • 10 team members import it simultaneously—everyone has identical configurations in 30 seconds
  • Company updates data governance rules, pushes new template to Git, everyone auto-imports on next launch
  • Consultant delivers project with template included for client team continuity

Configuration Transfer ensures zero-friction collaboration and instant standardization across your entire data organization.


Inline Visualizations in OpenAnalyst

Inline visualizations eliminate context switching.

OpenAnalyst connects directly to your code and renders charts in your workspace with no file exports.

With Inline Visualizations, you can:

  • Create charts instantly using ```chart code blocks that render like Mermaid diagrams
  • Iterate faster by seeing visualizations immediately after analysis runs
  • Stay in flow without switching between IDE, notebook, and image viewers
  • Share visual insights embedded directly in chat history
  • Support all major chart types: bar, line, pie, scatter, heatmaps

Use Cases:

  • A Data Analyst runs cohort analysis and creates a retention heatmap—rendered inline in 2 seconds
  • A Data Scientist compares model performance with side-by-side ROC curves in the chat window
  • A product team reviews A/B test results with bar charts generated on-demand during the conversation

Inline Visualizations make OpenAnalyst your fastest path from question to insight.


Task Orchestration in OpenAnalyst

Task Orchestration breaks complexity into manageable steps.

OpenAnalyst's Orchestrator mode lets you:

  • Decompose complex projects into subtasks assigned to specialized agents
  • Run parallel analyses with multiple agents working simultaneously
  • Track hierarchical workflows with visual subtask trees (@parent@child syntax)
  • Aggregate results automatically from all subtasks into final deliverable
  • Resume interrupted work from checkpoints without losing context

Use Cases:

  • Market analysis project: Research Analyst finds competitors → Data Analyst analyzes survey data → Business Analyst calculates ROI → All synthesized into executive summary
  • ML pipeline: Data Engineer builds ETL → Data Scientist trains model → Data Analyst creates evaluation dashboard
  • Quarterly reporting: 5 agents run parallel analyses on different metrics, Orchestrator compiles into single report

Task Orchestration turns OpenAnalyst into your AI project manager for data work.


Why OpenAnalyst is Different

Traditional Method

New employee joins data team
↓
Reads 50 pages of documentation
↓
Manually configures IDE, installs libraries
↓
Learns company conventions through trial and error
↓
Writes first useful analysis after 2-3 weeks

OpenAnalyst Method

New employee joins data team
↓
Imports company-analytics.json (30 seconds)
↓
Activates team template with 5 specialized agents
↓
AI follows company standards automatically
↓
Delivers first insight within 1 hour

Getting Started

  1. Install OpenAnalyst extension in VS Code
  2. Import your team's config (if available) or start fresh
  3. Activate a template from the example library or create your own
  4. Select an agent (Data Analyst recommended for first use)
  5. Ask your question in natural language and let OpenAnalyst work

Example first task:

"Load sales_data.csv, analyze Q4 trends by region, and create a bar chart comparing top 5 performing regions"

OpenAnalyst will:

  • Read your data file
  • Generate Python/pandas code
  • Run the analysis
  • Create an inline bar chart
  • Explain key findings

All in minutes, not hours.


Open Source & Extensible

OpenAnalyst is fully open source with:

  • No vendor lock-in (bring your own API keys)
  • Support for local models (Ollama, LMStudio)
  • MCP (Model Context Protocol) integration for custom tools
  • Active community contributions
  • Transparent codebase you can audit and extend

Built on the evolution of Cline → Roo Code → KiloCode, OpenAnalyst combines the best of each while pioneering template-driven collaboration for data teams.


Learn More


OpenAnalyst: Empowering data teams to deliver insights faster through AI collaboration, not replacement.