Not All Tokens Are Created Equal: Cutting the Cost of Agentic Code Search

Description

AI agents spend much of their time reading and searching code, consuming huge numbers of tokens in the process. So how can you make agentic coding workflows more efficient? Watch this on demand session from Cockroach Labs and Takara.ai to explore how smarter code retrieval can reduce token consumption and help bring down the cost of agentic development. The session looks at where tokens are being consumed, how CockroachDB can support a fresh, distributed vector index, and introduces Miru, Takara.ai’s open source, agent agnostic code search. You’ll also see Miru in action against a large codebase, with benchmark results against a grep baseline. Ideal for engineering leaders, platform teams and anyone building or managing AI powered development workflows.