Streaming SQL Version Control
Execute zero-copy, sub-millisecond time-travel SQL queries directly over immutable Apache Arrow Merkle DAG streams. Access repository state anywhere with zero database migrations.
⚡ Interactive SQL Console
Ready (0ms)
| agent_id | step | accuracy | latency_ms | status |
|---|---|---|---|---|
| agent-leader | 1 | 0.88 | 120.5 | passed |
| agent-coder | 1 | 0.91 | 45.2 | passed |
📜 VCS Commit DAG
Repo: aivcs/aivcsd-stream
d5422863fdb3 (HEAD)
2026-09-03 12:09:00 UTC
Run 2: prompt optimization + reviewer agent added
Parents: c40053f59f40 · Tables: agent_evals, ast_nodes (3 rows)
c40053f59f40
2026-09-03 12:09:00 UTC
Run 1: baseline multi-agent coding swarm
Roots: None · Tables: agent_evals, ast_nodes (2 rows)
💻 Quickstart Client Access
# Python / DuckDB / Polars (Access Anywhere)
import polars as pl
import pyarrow.flight as flight
client = flight.connect("grpc+tls://stream.aivcs.io:443")
reader = client.do_get(flight.Ticket("aivcs/aivcsd-stream@main"))
df = pl.from_arrow(reader.read_all())
print(df)