Coding Agent Swarms, Part 5: Running the Fleet From Your Phone
The Last Mile Is the Operator The first four parts of this series built the substrate: foundation, fleet, multi-fleet …
Chat Studio v1.1.0 is here. The first major release since launch, and it addresses the most-requested features head on: inline data visualization that does not require exporting to a spreadsheet, a query engine that is significantly smarter and faster, and a proper audit trail system that actually gives compliance teams what they need. This is the release that takes Chat Studio from a great prototype to a production-grade tool.
What is Chat Studio?
Chat Studio is a multi-model conversational AI interface that connects directly to your databases and documents. Ask questions in plain English, see the SQL it generated, get charts inline. Available as a standalone desktop app, embedded in AI Lab and IDE, or running in the browser workspace.
What is new in v1.1.0
Inline Charts and Data Visualization
This is the headliner. You can now ask Chat Studio to visualize query results directly in the conversation without exporting to another tool. Ask “show me monthly revenue as a bar chart” and get a fully rendered, interactive chart inline. Charts support bar, line, scatter, pie, and area formats. They can be exported as PNG or copied directly to clipboard.
Visualizations are generated from query results automatically when the response looks like tabular data. You can override the chart type or request raw table view at any time in the same thread.
Query Engine Overhaul
The underlying query engine has been rewritten for this release. It now maintains full schema context across multi-turn conversations, so you can reference “that table from earlier” or “the same filter we used last time” and get correct results. Join resolution across multiple schemas is significantly more reliable.
Query execution is also faster. Streaming result rendering means you start seeing data as soon as the database responds rather than waiting for the full result set to load.
SQL Transparency Panel
SQL Transparency was a feature in v1.0.0. In v1.1.0, it has been upgraded to a full panel. Every query Chat Studio runs is shown in a collapsible, syntax-highlighted SQL panel alongside the response. You can copy it, edit it, and rerun it directly. For data teams that need to know exactly what is hitting their databases, this is now genuinely useful rather than just a checkbox.
Audit Trail and Compliance Exports
Enterprise teams running Chat Studio under Zentinelle governance now get a complete audit trail: who asked what, against which data source, with the exact SQL generated, timestamped and signed. Audit logs are exportable in CSV and JSON formats directly from the admin panel.
Role-based access control has also been tightened. Admins can now restrict specific users or groups to read-only access on specific data sources without affecting their access to other connections.
Context Window Management
Long conversations with large result sets used to degrade in quality as the context window filled up. v1.1.0 introduces intelligent context compression that summarizes older query results while keeping the conversation coherent. You can now run hour-long data exploration sessions without the assistant losing track of what you established at the start.
Multi-Model Switching Mid-Conversation
You can now switch models mid-conversation without starting a new thread. Switch from Claude to GPT-4 when you want faster responses for simple lookups, then switch back to Claude for complex multi-table analysis. Conversation history is preserved across model switches.
Bug Fixes
- Fixed an issue where queries against MySQL databases with ENUM columns would return incorrect type information
- Fixed the “explain this query” feature producing responses that did not match the actual generated SQL
- Resolved a display bug where very wide result tables would overflow the conversation panel
- Fixed a crash when connecting to PostgreSQL instances with schemas containing circular foreign key references
- Fixed role-based access restrictions not applying correctly to users added after initial setup
Download
Free for personal use. Bring your own API keys for any supported model provider. Enterprise governance and audit trail features available with Zentinelle.
What is Next
v1.2.0 will introduce scheduled query reports (run a query on a schedule, deliver results to Slack or email), document upload support for asking questions against PDFs and spreadsheets, and expanded visualization types including pivot tables and heatmaps.
