Memory API
Store and search long-term memory over one API instead of wiring retrieval logic into every agent flow separately.
Memory infrastructure for production AI
Persistent memory for AI products that need real controls.
Neuralbase gives your app long-term memory, semantic retrieval, document parsing, usage enforcement, and paid backups in one platform. Ship recall without building the operational layer yourself.
Persistent storage on every plan
Per-user collection isolation
Free hard stops, paid soft warnings
Manual restore for paid snapshots
Quickstartlive API
curl -X POST https://api.neuralbase.cloud/v1/memories \
-H "Authorization: Bearer nb_live_..." \
-d '{
"userId": "u_173",
"content": "Prefers short answers"
}'Memory hit
Prefers concise responses
Source: user chat · confidence high
Usage state
340 / 1,000 API calls
Free plan hard-stops at monthly limit
Backups
Daily on Pro and Enterprise
Persistent storage remains on every plan
Works with your stack
OpenAILangGraphCrewAIQdrantVoyage AISupabaseRedisBullMQHonoVercel
What ships in the box
The product is more than vector storage.
Neuralbase is the memory layer plus the operational surface around it: document flows, quotas, paid backups, and account-level isolation.
Document parsing
Parse files, inspect extracted output, and decide whether document results should become durable memory later.
Usage and quotas
Meter API calls, document parses, and stored memories with plan-aware limits, warnings, and rate controls.
Paid backups
Persistent storage exists for every plan. Daily snapshots with retention are reserved for paid accounts.
Built around constraints
Designed for an actual runtime, not a toy demo.
The platform has already been forced to care about billing, rate limits, isolation boundaries, persistence, and backup policy. That makes the product sharper than a memory wrapper that only shows the happy path.
Built for real constraints
This platform is shaped around an actual single-VM deployment, which forces the product to care about quotas, rate limits, storage boundaries, and backups early.
Operational surface included
Usage, backup state, restore requests, and plan enforcement are part of the product surface, not hidden side effects in the backend.
Isolation is explicit
Data is no longer sitting in one shared vector collection. Each account writes to its own collection boundary.
Use cases
Built for products that need context to persist.
Support and success copilots
Carry customer preferences, prior resolutions, and account context into the next reply without replaying every historical message.
Document-heavy internal tools
Parse files, extract structured text, and keep the important parts searchable for later agent workflows and internal search.
Sales and CRM assistants
Remember objections, next steps, and stakeholder details across long cycles so the assistant stays grounded between calls.
Research and analyst workflows
Combine parsed document context with durable memory to keep working context alive across multi-step research tasks.
Pricing
Start free. Upgrade when the workload earns it.
The pricing model reflects actual resource pressure: API traffic, stored memories, document parsing, rate limits, and paid snapshots.
Free
$0
For first production trials
Validate memory value before paying for higher throughput or scheduled backups.
1,000 API calls per month
500 stored memories
5 document parses per month
20 requests per minute
Persistent storage
Pro
$49
Per workspace / month
For customer-facing workloads that need higher throughput and paid operational protection.
50,000 API calls per month
25,000 stored memories
200 document parses per month
200 requests per minute
Daily backups with 30-day retention
Enterprise
Custom
Enterprise
For negotiated limits, procurement review, and platform coordination.
Custom API, storage, and parse limits
Negotiated rate limits
Daily backups with 30-day retention
Manual restore coordination
Commercial support path
FAQ
Common questions.
Recent work
What has shipped recently.
Get started
Give your AI product memory without ignoring the operational reality.
Start on the free plan, wire the API into a real workflow, and upgrade when you need higher limits and scheduled backups.