Analyzes Convex function read patterns to project spend and recommend cost fixes.
npx skills add https://github.com/openclaw/clawhub --skill convex-cost
Please help me install this Agent Skill. Skill: convex-cost Source: https://github.com/openclaw/clawhub/tree/main/.agents/skills/convex-cost Command: npx skills add https://github.com/openclaw/clawhub --skill convex-cost
Convex bills on function calls and bandwidth, so a few functions reading far more data than expected can quietly drive up spend. This Skill makes that spend legible by attributing bytes/documents-read evidence to the functions responsible, projecting how costs grow with traffic, and naming the cheapest fix for each driver.
.collect() grows linearly with table size while an indexed .take(n) stays flat) instead of inventing false-precision dollar figures..withIndex, replacing .collect() with .paginate/.take, or caching hot reads, and emits cost-class findings for convex-expert/convex-advisor.messages:list reads the whole table every call and that indexing it drops the cost roughly 100x.Ask the agent to preview what my Convex deployment will cost and rank the functions driving the spend.
High-intent search queries and answers about installing and using this skill.
Estimate Convex cost by ranking functions by bytes or documents read per call multiplied by call volume, using the deployment's insights data. With no traffic yet, estimate from query shapes instead, since a full-table scan on a growing table is a future cost even at zero usage.
Reduce Convex read costs by replacing full-table `.collect()` scans with indexed `.withIndex` queries, using `.paginate` or `.take` instead of reading entire tables, adding aggregate components for counts, and caching hot reads. Each fix targets the specific function driving the spend.
Yes, but differently. Without usage data, the analysis prices query shapes instead of measured reads: a `.collect()` on a table projected to grow is flagged as a future linear cost, while indexed access patterns are identified as flat regardless of table size.
Cost equals data read per call times call volume, so a cheap-per-call function invoked constantly can outweigh an expensive rare one. Both factors must be examined together to find the true cost drivers.
Projections give growth curves and relative magnitudes, not exact dollar figures, because Convex pricing depends on plan and changes over time. The insights data is also cloud and user-auth only, so preview deployments cannot be analyzed.
matixlolQuery Convex databases via the Convex MCP server for schema and data inspection.
davila7Implements Convex backends with schemas, TypeScript functions, real-time subscriptions, and deployment workflows.
kollaboraiAnalyze EXPLAIN plans and execution times to optimize SQL queries.
OpenRouterTeamAnswer natural language questions about OpenRouter usage and billing data.