Improve GitHub discovery docs

.github/ISSUE_TEMPLATE/config.yml

@@ -1 +1,5 @@
 blank_issues_enabled: false
+contact_links:
+  - name: Questions and usage help
+    url: https://github.com/intraq-dev-ai/intraq/discussions
+    about: Ask setup, usage, and roadmap questions in GitHub Discussions.

.github/ISSUE_TEMPLATE/connector_request.yml

@@ -0,0 +1,61 @@
+name: Connector request
+description: Request or help implement a data-source connector.
+title: "[Connector]: "
+labels:
+  - enhancement
+  - type:connector
+  - area:data-sources
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Use this for public-source data-source connectors only. Do not include credentials, private endpoint URLs, customer exports, or non-public schemas.
+  - type: input
+    id: connector
+    attributes:
+      label: Connector
+      description: Which database, warehouse, file store, or API should IntraQ support?
+      placeholder: DuckDB, BigQuery, Snowflake, Athena, Trino, GCS, etc.
+    validations:
+      required: true
+  - type: dropdown
+    id: connector_type
+    attributes:
+      label: Connector type
+      options:
+        - SQL database / warehouse
+        - Object storage / files
+        - API / HTTP
+        - Other
+    validations:
+      required: true
+  - type: textarea
+    id: use_case
+    attributes:
+      label: Use case
+      description: What kind of reporting or dashboard workflow should this unlock?
+    validations:
+      required: true
+  - type: textarea
+    id: config
+    attributes:
+      label: Expected configuration
+      description: List only non-secret field names, not real values.
+      placeholder: |
+        host
+        port
+        database
+        schema
+        auth method
+  - type: textarea
+    id: acceptance
+    attributes:
+      label: Acceptance criteria
+      description: What should work when this connector is complete?
+      placeholder: |
+        - Connection test
+        - Schema discovery
+        - SQL Editor query preview
+        - Analyzer
+        - Dashboard Builder
+

README.md

@@ -5,7 +5,7 @@
 [![Docker](https://img.shields.io/badge/docker-compose-2496ED?logo=docker&logoColor=white)](docker-compose.yml)
 [![Node](https://img.shields.io/badge/node-24.x-339933?logo=node.js&logoColor=white)](.nvmrc)
 
-[Website](https://intraq.dev) · [Docs](docs/DEMO_GUIDE.md) · [Quickstart](QUICKSTART.md) · [Configuration](docs/CONFIGURATION.md) · [Contributing](CONTRIBUTING.md)
+[Website](https://intraq.dev) · [Docs](docs/DEMO_GUIDE.md) · [Quickstart](QUICKSTART.md) · [Configuration](docs/CONFIGURATION.md) · [Comparisons](docs/comparisons/README.md) · [Discussions](https://github.com/intraq-dev-ai/intraq/discussions) · [Contributing](CONTRIBUTING.md)
 
 **IntraQ is a source-available AI business intelligence platform that lets teams query SQL-backed operational data using natural language, generate trusted SQL, and build interactive dashboards without traditional BI complexity.**
 
@@ -92,6 +92,13 @@ Explore practical question sets in [`examples/`](examples/README.md).
 
 intraQ is not trying to replace every enterprise reporting suite. It is focused on operational BI workflows where teams want AI-assisted SQL, local control, and dashboards from governed data models.
 
+Detailed comparison pages:
+
+- [IntraQ vs Metabase](docs/comparisons/intraq-vs-metabase.md)
+- [IntraQ vs Wren AI](docs/comparisons/intraq-vs-wren-ai.md)
+- [IntraQ vs Power BI](docs/comparisons/intraq-vs-power-bi.md)
+- [IntraQ vs Lightdash](docs/comparisons/intraq-vs-lightdash.md)
+
 ### intraQ vs Power BI
 
 Power BI is a broad reporting suite. intraQ focuses on natural language SQL, evidence-backed AI answers, and turning operational questions into live dashboard components from SQL-backed models.
@@ -168,6 +175,7 @@ Focused workflow docs:
 - [SQL Editor](docs/SQL_EDITOR.md)
 - [MCP tools](docs/MCP.md)
 - [AI provider setup](docs/AI_PROVIDER_SETUP.md)
+- [Comparisons](docs/comparisons/README.md)
 - [Publication checklist](docs/PUBLICATION_CHECKLIST.md)
 
 ## License and public source scope

ROADMAP.md

@@ -10,19 +10,22 @@ This roadmap is intentionally focused on the public source package.
 - Generic sample sales dataset.
 - Metadata-grounded Analyzer and Dashboard Builder.
 - MCP tools for local product workflows.
+- Public GitHub issue labels, discussions, templates, and connector roadmap.
 
 ## Next
 
+- Fix remaining web type-check debt so the web workspace can become a required CI check.
+- Add a real demo GIF showing: ask question → SQL/evidence → chart → saved dashboard component.
+- Add smaller `good first issue` tasks around connector work.
 - More smoke and regression tests for Analyzer, SQL, dashboards, and setup.
 - Public sample question packs with expected SQL/results.
 - Better first-run onboarding.
-- Smaller web bundles through route-level code splitting.
-- More connector documentation.
+- More connector documentation and connector test fixtures.
 - Clearer evidence surfaces for Analyzer and generated dashboards.
 
 ## Later
 
-- Public-source connector examples.
+- Public-source connector examples and runtimes, starting with DuckDB, BigQuery, and Snowflake.
 - More generic sample datasets.
 - Public benchmark reports for sample datasets.
 - Import/export for metadata models.

docs/comparisons/README.md

@@ -0,0 +1,11 @@
+# Comparisons
+
+These pages explain where IntraQ fits compared with common BI and AI analytics tools.
+
+The intent is practical positioning, not claiming that one tool is universally better.
+
+- [IntraQ vs Metabase](intraq-vs-metabase.md)
+- [IntraQ vs Wren AI](intraq-vs-wren-ai.md)
+- [IntraQ vs Power BI](intraq-vs-power-bi.md)
+- [IntraQ vs Lightdash](intraq-vs-lightdash.md)
+

docs/comparisons/intraq-vs-lightdash.md

@@ -0,0 +1,34 @@
+# IntraQ vs Lightdash
+
+Lightdash is a BI tool built around dbt projects and metrics defined in code.
+
+IntraQ is focused on AI-assisted operational BI over SQL-backed data sources, with Analyzer, SQL Editor, Dashboard Builder, and metadata-guided answer generation.
+
+## Where Lightdash is strong
+
+- dbt-native analytics workflow.
+- Metrics and dimensions defined close to transformation code.
+- Strong fit for analytics engineering teams.
+- Good version-control alignment when the organization already uses dbt.
+
+## Where IntraQ is different
+
+- Does not require a dbt-first workflow.
+- Includes an AI Analyzer and dashboard-builder workflow.
+- Designed for operational teams working from existing SQL data models, metadata, and dictionaries.
+- Emphasizes evidence-backed AI answers and live dashboard components.
+- Includes source-available self-hosted setup with local AI provider configuration.
+
+## Choose IntraQ when
+
+- You want natural language SQL and AI-generated dashboards over operational data.
+- You do not have a mature dbt workflow.
+- You need business users to move from question to dashboard component quickly.
+- You want AI responses to expose assumptions, source context, and SQL evidence.
+
+## Choose Lightdash when
+
+- Your analytics layer is already built in dbt.
+- Your team wants BI tightly coupled to analytics engineering workflows.
+- You prioritize metrics-as-code over AI-assisted dashboard creation.
+

docs/comparisons/intraq-vs-metabase.md

@@ -0,0 +1,34 @@
+# IntraQ vs Metabase
+
+Metabase is a mature self-service BI tool for querying databases, creating dashboards, and helping teams explore data.
+
+IntraQ is focused on AI-assisted operational BI: natural language SQL, evidence-backed answers, dashboard generation, and reusable SQL-backed components.
+
+## Where Metabase is strong
+
+- Mature query builder and dashboard workflows.
+- Broad database support.
+- Familiar self-service BI experience.
+- Good fit for teams that primarily want classic dashboarding and exploration.
+
+## Where IntraQ is different
+
+- AI Analyzer is built into the workflow.
+- Answers are expected to stay tied to SQL, selected data source, result rows, and assumptions.
+- Dashboard Builder can turn AI answers into live dashboard components.
+- Metadata, dictionaries, saved SQL models, relationships, and dashboard context guide AI behavior.
+- Public source focuses on self-hosted operational analytics and local AI provider configuration.
+
+## Choose IntraQ when
+
+- You want an AI dashboard builder over SQL-backed operational data.
+- You want natural language SQL with visible evidence and assumptions.
+- You want to capture reusable dashboards from AI-assisted analysis.
+- You are building embedded or internal reporting workflows around operational metrics.
+
+## Choose Metabase when
+
+- You mainly need a mature classic BI tool.
+- Broad connector maturity matters more than AI-assisted answer verification.
+- Your users prefer visual query building over AI-guided analysis.
+

docs/comparisons/intraq-vs-power-bi.md

@@ -0,0 +1,34 @@
+# IntraQ vs Power BI
+
+Power BI is a broad enterprise analytics suite with a large ecosystem, Microsoft integration, modeling tools, and mature visualization features.
+
+IntraQ is narrower: self-hosted AI business intelligence for operational SQL-backed data, natural language SQL, evidence-backed answers, and AI-created dashboard components.
+
+## Where Power BI is strong
+
+- Enterprise reporting and governance.
+- Rich visualization and modeling ecosystem.
+- Microsoft 365, Fabric, and Azure integration.
+- Large partner and consultant ecosystem.
+
+## Where IntraQ is different
+
+- Designed for source-available, self-hosted deployment.
+- Focused on SQL-backed operational analytics.
+- AI answers are expected to expose SQL, evidence, and assumptions.
+- Dashboard components can be created from AI-assisted analysis.
+- Local/admin-configured AI providers are part of the public app setup.
+
+## Choose IntraQ when
+
+- You want a lightweight operational BI workflow.
+- You want AI-assisted SQL and dashboards without committing to a full enterprise BI suite.
+- You want to inspect and self-host the source.
+- You need product/internal reporting that can be embedded or automated.
+
+## Choose Power BI when
+
+- You need enterprise Microsoft ecosystem integration.
+- Your organization already standardizes on Power BI.
+- You need advanced enterprise governance and visualization breadth today.
+

docs/comparisons/intraq-vs-wren-ai.md

@@ -0,0 +1,47 @@
+# IntraQ vs Wren AI
+
+Wren AI focuses on natural language querying over a semantic/context layer and has broad connector ambitions.
+
+IntraQ focuses on verified operational BI: AI-assisted SQL, evidence-backed answers, dashboard creation, and reusable reporting workflows inside one self-hosted application.
+
+## Where Wren AI is strong
+
+- Natural language querying over modeled data.
+- Connector breadth is a major product direction.
+- Strong positioning around AI-native data access.
+
+## Where IntraQ is different
+
+- Combines Analyzer, SQL Editor, Dashboard Builder, data-source management, and MCP workflows in one app.
+- Emphasizes answer evidence: SQL, selected source/model, row results, assumptions, and missing-context warnings.
+- Turns AI-assisted answers into live dashboard components.
+- Includes operational examples for hospitality/POS, energy retail, ecommerce, SaaS, and Postgres dashboards.
+- Uses local/admin-configured AI providers rather than an IntraQ managed AI dependency in the public source.
+
+## Connector roadmap
+
+IntraQ already tracks connector parity work as public issues:
+
+- DuckDB
+- BigQuery
+- Snowflake
+- Redshift
+- Oracle
+- Athena
+- Trino/Presto
+- Spark/EMR
+- Doris
+- GCS
+- MinIO/S3-compatible storage
+
+## Choose IntraQ when
+
+- You want AI BI plus dashboard generation, not only natural language querying.
+- You want evidence-backed operational dashboards.
+- You want a source-available app that can be self-hosted and inspected.
+
+## Choose Wren AI when
+
+- Your top priority is semantic-layer-first natural language querying.
+- You need one of its currently supported connectors before IntraQ supports it.
+