Perplexity open-sources CobbleDB, slashing latency and using AI agents for development · 7 sources tracked

By PulseAugur Editorial · [7 sources] · 2026-09-15 20:16

Perplexity has developed and is open-sourcing CobbleDB, a new key-value database designed for efficient batch reads, significantly improving their search workload performance. This new system replaced Amazon DynamoDB, resulting in a dramatic reduction in median batch-read latency from 31.4 ms to 5.60 ms and a p99 latency drop from 123 ms to 24.2 ms. The development of CobbleDB was accelerated by a team of AI agents, with the core infrastructure built in two months by two engineers and hundreds of these agents. AI

IMPACT Accelerates AI search performance and offers a new tool for developers, potentially improving efficiency in AI-related applications.

RANK_REASON Perplexity is releasing a new database technology and detailing its development, which is a product launch/release.

Read on X — Perplexity →

AI-generated summary · Google Gemini · from 7 sources. How we write summaries →

Perplexity open-sources CobbleDB, slashing latency and using AI agents for development · 7 sources tracked

How we ranked this

Signal score: 7 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; Perplexity is releasing a new database technology and detailing its development, which is a product launch/release.
Source corroboration: 7 independent sources; Strong cross-source corroboration — multiple independent publishers covered this within the clustering window.
Topics: infra, product; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [7]

X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-09-15 20:17

CobbleDB is purpose-built for repeated batch reads of prepared page records, optimizing performance and cost for Perplexity’s search workload.

CobbleDB is purpose-built for repeated batch reads of prepared page records, optimizing performance and cost for Perplexity’s search workload. We plan to open-source CobbleDB for other teams building AI search. Read more: https://t.co/iK1dd8WDhR
X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-09-15 20:17

After replacing DynamoDB, production before-and-after measurements show median batch-read latency falling from 31.4 to 5.60 ms and p99 from 123 to 24.2 ms.

After replacing DynamoDB, production before-and-after measurements show median batch-read latency falling from 31.4 to 5.60 ms and p99 from 123 to 24.2 ms. Our internal cost model estimates at least 20% savings relative to DynamoDB. https://t.co/Bs8c9vsRte
X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-09-15 20:17

That separation lets each replica ingest updates at its own pace.

That separation lets each replica ingest updates at its own pace. Pillar stores versioned document state in YTsaurus and queues exports. Lorry writes partitioned S3 batches that CobbleDB replicas apply independently and in order, so recovering nodes don’t delay others. https://t…
X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-09-15 20:17

Optimizing reads also required changing how updates reached the store.

Optimizing reads also required changing how updates reached the store. We separated durable document state, update delivery, and serving across Pillar, Lorry, and CobbleDB. https://t.co/pL7vCBTyC3
X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-09-15 20:16

The architecture prioritizes fast batch reads. CobbleDB groups page keys by partition and reads them in parallel through RocksDB MultiGet.

The architecture prioritizes fast batch reads. CobbleDB groups page keys by partition and reads them in parallel through RocksDB MultiGet. The router favors same-zone replicas and queries another copy when a replica is slow, so one delayed response need not hold up the batch. ht…
X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-09-15 20:16

We accelerated CobbleDB development with a team of proactive coding agents.

We accelerated CobbleDB development with a team of proactive coding agents. The agents provided continuous inspection and follow-through across sessions under the direction of engineers responsible for architecture, review, and production authorization.
X — Perplexity TIER_1 English(EN) · perplexity_ai · 2026-09-15 20:16

We’re publishing research on how we built CobbleDB, our key-value database that serves web content for Perplexity search.

We’re publishing research on how we built CobbleDB, our key-value database that serves web content for Perplexity search. Two engineers and a team of hundreds of proactive, always-on AI agents built the core infrastructure in two months. https://t.co/OsCtY15UW6

How we ranked this

COVERAGE [7]

CobbleDB is purpose-built for repeated batch reads of prepared page records, optimizing performance and cost for Perplexity’s search workload.

After replacing DynamoDB, production before-and-after measurements show median batch-read latency falling from 31.4 to 5.60 ms and p99 from 123 to 24.2 ms.

That separation lets each replica ingest updates at its own pace.

Optimizing reads also required changing how updates reached the store.

The architecture prioritizes fast batch reads. CobbleDB groups page keys by partition and reads them in parallel through RocksDB MultiGet.

We accelerated CobbleDB development with a team of proactive coding agents.

We’re publishing research on how we built CobbleDB, our key-value database that serves web content for Perplexity search.

RELATED ENTITIES

RELATED TOPICS