January was all about visibility. Custom observability dashboards let you build the monitoring views you actually need, model pricing gives you per-run cost estimates, and a round of UI improvements makes everyday workflows faster.
Custom Observability Dashboards
You can now build fully custom dashboards to monitor your agents exactly the way you want. Create multiple dashboards per project, add widgets, rearrange them with drag-and-drop, and adjust the time window with a flexible date range picker.
Four widget types are available out of the box:
- ✓Area charts: Visualize runs, token usage, latency, or cost over time with grouping by agent, status, or model
- ✓Horizontal bar charts: Compare metrics across agents, models, or connectors at a glance
- ✓Stat cards: Single-number KPIs like total runs, average latency, success rate, or estimated cost
- ✓Logs list: A live feed of recent agent runs with status, duration, and error details
Every widget is fully configurable. Pick a metric, choose how to aggregate it, filter by agent or connector, and select the visualization that makes the data clearest. Dashboards are saved per project and accessible from the new Observability tab in your project sidebar.
Model Pricing & Cost Tracking
Know what your agents cost. Connic now ships with pricing data for popular models from OpenAI, Anthropic, Google, and others. Each run shows an estimated cost based on the tokens it consumed and the model it used.
Pricing covers four token categories:
- •Input tokens: Prompt and context sent to the model
- •Output tokens: Generated response from the model
- •Cached input tokens: Tokens served from cache at a reduced rate
- •Thinking tokens: Reasoning tokens used by models with extended thinking
Global defaults work automatically, but you can override pricing for any model in your project under Settings > Observability. This is especially useful when you use fine-tuned models or negotiate custom rates with a provider. Cost data also feeds into the new dashboard widgets, so you can chart spending over time or compare costs across agents.
Dashboard UI Improvements
Several pages received usability improvements this month:
- •Connector overview: The connector list now shows live stats per connector, including total runs, success rate, and when it was last triggered
- •Connector detail: Expanded view with a configurable time range for run statistics and a cleaner layout for linked agents
- •Runs table: Now includes an estimated cost column so you can spot expensive runs at a glance
- •Pricing page: Redesigned feature comparison with expandable sections for Features, Support & Enterprise, and Compliance
- •404 page: A proper not-found page so dead links no longer land on a blank screen
llms.txt
We now publish a llms.txt file at the root of our website. This follows the emerging llms.txt standard and gives AI assistants a structured overview of Connic, including links to our documentation, connector guides, feature pages, and more.
If you use Cursor, Claude, ChatGPT, or other AI coding assistants, they can now understand what Connic does and reference the right documentation when helping you build and configure agents.