We are updating Backboard’s pricing to make costs predictable, usage visible, and comparisons fair.

This change reflects how developers actually build and scale stateful AI systems.

What’s Changing

Backboard now uses a simple, transparent pricing model:

• $9 per month subscription

• Usage-based fees

• A Free tier with $5 in credits

No tiers to decode. No bundled plans. No surprise minimums.

Free Tier

We have added a Free tier that includes up to $5 in credits.

These credits can be used across any Backboard service, including:

• Memory reads

• Memory writes

• Stored memories

• Tokens

This tier is designed for experimentation, evaluation, and early development without requiring a paid commitment.

Why We Made This Change

As more teams shipped production workloads on Backboard, two patterns became clear.

First, developers wanted a low, predictable entry point. Something easy to approve, easy to explain, and easy to start with.

Second, usage varies significantly. Some applications store minimal state. Others run long-lived, high-throughput systems with frequent context retrieval, persistence, and memory growth. Flat pricing either penalizes small teams or breaks under scale.

Usage-based pricing keeps this proportional and fair.

How Usage is Priced

Usage is billed based on what the system actually does.

• Memory reads are priced at $0.003 per read
  Reads typically occur when context or state is retrieved during execution.

• Memory writes range from $0.0016 to $0.005 per write
  Writes occur when new state, memory, or derived context is persisted. We batch when possible saving you money

• Stored memories are priced at $0.25 per 100,000 stored memories
  This reflects the cost of long-term storage, indexing, and retrieval readiness.

• Tokens are billed at the underlying rates set by the model or network providers you choose

We do not arbitrarily mark up token pricing. As we introduce platform-level efficiencies, we will reflect those improvements directly in usage pricing rather than introducing new tiers or fees.

Visibility and Cost Monitoring

Usage and charges are always visible.

You can see:

• How much you have used

• What you are being charged for

• How costs break down across reads, writes, stored memories, and tokens

Cost visibility is built into the platform and does not require support tickets or manual reporting.

What You are Actually Paying For

Backboard is not a standalone memory database.

The API includes:

• Stateful threads

• Memory reads and writes

• Retrieval-augmented generation

• Orchestration

• LLM routing across providers

• Execution and lifecycle management

When comparing pricing, it is important to compare full systems. Backboard replaces multiple layers of a typical AI stack and exposes them through a single API.

Backboard should not be compared to standalone vector databases or raw token proxies.

What Stays The Same

• No long-term contracts

• No forced tiers

• No hidden minimums

You pay a small monthly subscription to access the platform, then pay for what you use.

Why This Model Works Long Term

In the short term, this keeps the barrier to entry low for developers experimenting or shipping early products.

In the long term, it aligns incentives. If we make the system more efficient, your costs go down. If your product succeeds and usage grows, pricing scales proportionally instead of jumping tiers.

That alignment is intentional.