Every answer,
grounded in now.
Feed your models the present, not last night's batch. Basalt turns orders, tickets and sessions into typed context that agents read in milliseconds.
A model is only as current as its context.
Most agents read from a warehouse that refreshes overnight. They answer with confidence about orders that already shipped and tickets that already closed.
of agent errors trace back to stale or missing context.
median gap between an event and the warehouse table a model reads.
p99 to fetch a full context bundle, joined and policy-checked.
From event to answer in one hop.
No nightly export, no cache to warm, no second copy to keep in sync. The row your app just wrote is the row your agent reads.
Events land.
Orders, tickets, clicks and payments stream in from Kafka, Postgres CDC, webhooks and the SDK. Each event counts once.
Views stay current.
SQL views join and window those events as they arrive, so every customer has one context row that moves with them.
Context is served.
Agents call one endpoint over MCP, REST or SQL. Row policies and PII masking run before anything leaves the region.
The model answers.
The agent gets typed facts with an as_of stamp and cites them. No nightly job sits between the event and the reply.
See every read an agent makes.
Each tool call is traced with its query, the rows it touched and the policy that allowed it. When an answer looks wrong, replay the exact context the model saw.
- Latency, rows and policy decision for every call
- Replay any turn with time travel on every view
- Export to OpenTelemetry, Spanlog or Watchly
Four milliseconds or eleven hours.
We timed one context request three ways: Basalt, a vector search followed by a warehouse join, and the nightly batch most teams start with.
Bring any model. Keep your stack.
Basalt speaks MCP, standard tool schemas, REST and Postgres. Plug it into the model and framework you already run.
Nimbia
Corvid AI
Zephyrine
Ocelot
Chainweave
Indexly
Driftcloud AI SDK
Pinevale
Agents that act on what just happened.
Answer with the order that shipped a minute ago.
Copilots read the customer, their open tickets and the latest shipment as one typed bundle. Replies quote facts from this minute, and every read follows the agent’s row policy.
- One bundle per customer: profile, orders, tickets
- PII masked before it reaches the prompt
- Citations link back to the source event
Give fraud agents the last thirty seconds.
A fraud agent sees the card, the device and every attempt in the current session before it decides. Velocity features update on each attempt, so the score reflects what just happened.
- Velocity features updated on every attempt
- Device and session joins inside one view
- Exact replay of any decision for reviewers
Recommend from what is in the cart right now.
Ranking models read live stock, price and the shopper’s current session. A sold-out item never reaches the shelf, and each new click changes the next suggestion.
- Session features refreshed on every click
- Live availability per SKU and region
- Embeddings stored beside the rows they describe
Ops agents that see the incident as it starts.
Ops agents read live metrics, deploys and on-call state from one place. They find the failing region, check the last change and page the right person with the evidence attached.
- Metrics, deploys and alerts joined by service
- Windows over the last five minutes, per region
- Actions gated by the same row policies
“Our copilot used to apologise for orders it could not see. Now it reads the same rows our support team does, a few milliseconds after they change. Escalations fell by a third in the first month.”
Questions AI teams ask first.
Keep one if you search millions of documents. Most agent context is structured: accounts, orders, tickets. Basalt serves those rows by key in milliseconds and can store embeddings beside them, so facts and retrieval come back in one call.
Views update within 50 ms of the source event at p99. Every response carries an as_of timestamp, and you can set a maximum age per tool so an agent declines to answer from data older than you allow.
Mark columns as sensitive in the view and choose mask, hash or drop for each API key. Redaction runs inside the engine before a bundle leaves the region, and every read is logged with the policy that applied.
Each agent key gets its own limit, from 100 to 50,000 reads a second, with short bursts allowed. If an agent loops, Basalt throttles that key and leaves your app traffic untouched.
Context reads are included in every plan. You pay for events ingested: Hobby is free up to 10M a month and Scale is $249 with 1B included. Enterprise adds reserved read capacity for large agent fleets.
Set your data in stone.
Free up to 10M events a month. No card, no sales call. Your first pipeline is live in about four minutes.