Install the CLI$ curl -fsSL https://rawtree.com/install.sh | bashSign in$ rtree loginInsert data$ rtree insert --table events --data '{"action":"click","user":"alice","value":42}'Query your data$ rtree query "SELECT action, count() FROM events GROUP BY action"
Built for agents (and humans) RawTree removes all database work between raw data and production queries.
Skip the upfront schema work without sacrifying performance
Send events, JSON, protobuf, Parquet/CSV/TSV files, or any type of telemetry directly. RawTree creates dynamic columns on ingest, so tables evolve with your applications. No schema, migration, or modeling work upfront.
Purpose-built for agentic coding and observability
RawTree exposes primitives agents already understand. Use the API, SDKs, CLI, or MCP server to let agents build analytic features, analyze logs and traces, debug systems or track their own work.
Performance that adapts to your production workloads
RawTree infers compression at ingestion time, data types at query time and optimizes recurring access patterns automatically with JIT projections. Optimization follows usage, no guesswork involved.
Use cases A single analytical layer for telemetry, events and agent data.
Send traces, logs, metrics, and wide events to RawTree through native OTLP endpoints. Query telemetry with SQL or PromQL-compatible queries without modeling every signal upfront.
import{RawTree}from"@rawtree/sdk";constrawtree=newRawTree({apiKey:process.env.RAWTREE_API_KEY!,});awaitrawtree.insert("events", [{event:"signup",user_id:"u_123"},]);constresult=awaitrawtree.query<{total:number}>("SELECT count() AS total FROM events GROUP BY event");
Real-time analytics
RawTree exposes primitives agents already understand. Stream raw application events into RawTree and query them as soon as they land. Build analytics over changing event payloads at workloads.
Capture tool calls, commands, traces, logs, and runtime events from agent runs. Reconstruct what happened with SQL or let an agent inspect its own execution history.
Run RawTree Cloud with autoscaling or fine control over your clusters. Multi-cluster deployments and multi-database setups. Or run it on your own infrastructure with data-plane control.
Official RawTree Terraform provider to wire AWS sources into RawTree. Kinesis, Firehose, CloudWatch, AWS WAF, CloudTrail, and more. Land operational data directly into queryable tables.
See RawTree query behavior, ingestion health, and performance signals without any extra tooling. Available via the UI, CLI, MCP or APIs
Collaborate with your team
Share data across teams and environments. Enable your agents with simple API keys.
Messy JSON? Contribute a 'transform'
Query JSON nested data naturally. But if it gets too complex, contribute transformation code to shape it.
"It's the S3 of the analytics category. Ingest and query. It just works"
Guillermo Rauch
CEO, Vercel
"RawTree let us complete this in hours, with confidence it will keep working as Kiro’s data evolves and new sources are added."
Ricardo Edo
CTO at Galeo
"It works especially well because it allows you to quickly move from a legacy, heterogeneous setup to a standardized core that can be dynamically adapted by the organization."
FJ
Javier Perez Ferrer
Data at NTT data
"It's just fast and simple. Works out of the box with coding agents. Exactly how it should be."
Rafa Moreno
Lead developer
Ready to get started
Start querying JSON with SQL in minutes. No credit card required.