Hebbrix: Pricing, Features & Best Use Cases

Our take: Hebbrix fits developers who want an OpenAI-compatible memory layer for production agents, starting free with 1,000 credits per month and paid plans from $19/month.

Main limitation: Usage is metered in credits, and on Free and Starter memory writes pause once the quota is exceeded.

By the AI Tool Atlas editors. Based on the vendor's official documentation; we have not tested this product hands-on. What “documented overview” means

Is Hebbrix right for you?

Hebbrix fits developers who want an OpenAI-compatible memory layer for production agents, starting free with 1,000 credits per month and paid plans from $19/month. Based on the vendor's documentation, not on a hands-on test.

Best for: Developers building AI agents

Also a good fit for

Teams running agents in production

Consider another tool if usage is metered in credits with documented weights

Or if

On Free and Starter, memory writes pause when quota is exceeded
Outcome Memory API listed from the Pro plan

Hebbrix pricing and usage limits

Prices checked Sep 29, 2026 on the official US pricing page. Official pricing ↗ Prices and limits change often.

The official Hebbrix pricing page as we saw it. hebbrix.com, captured Oct 1, 2026

Free access and trial conditions

Free plan with 1,000 credits per month, no credit card.

No separate trial; Free plan serves as the entry point.

Paid plans: what changes as you upgrade

Free api

Free source (opens in a new tab)

Month

For: Trying it out and small projects

Limits

1,000 credits per month

Includes

Adaptive hybrid search
Profile extraction

1 more

Community support

API included

Starter api

$19/mo source (opens in a new tab)

Month

For: Shipping a first product

Limits

25,000 credits per month
Writes pause over quota

Includes

Everything in Free
Knowledge graph

1 more

Bring your own model key

API included

Pro api

$99/mo source (opens in a new tab)

Month

For: Agents running in production

Limits

200,000 credits per month

Includes

Everything in Starter
Outcome Memory API

2 more

Advanced analytics
Overage billing

API included

Scale api

$399/mo source (opens in a new tab)

Month

For: High-volume workloads

Limits

1,000,000 credits per month

Includes

Everything in Pro
Higher rate limits

1 more

Overage billing

API included

Enterprise api

Custom pricing source (opens in a new tab)

Custom

For: Teams with scale and compliance needs

Limits

Pay as you go

Includes

Bring your own keys, any model
Single sign-on

2 more

Security questionnaire and procurement review
Solutions engineering

API included

Usage caps, credits and extra costs

Overage billing can be enabled on Pro and Scale; managed-model calls also report token usage.

Not stated.

What is Hebbrix, and how does it work?

Hebbrix is a memory infrastructure service for AI agents and applications. It stores memories and a knowledge graph, and ranks recall using semantic, lexical, graph, temporal and reported-outcome signals. Developers can point an existing OpenAI client at the Hebbrix API base URL so memory is retrieved and injected automatically. The docs also cover an MCP server, SDKs, webhooks and multi-tenancy.

Hebbrix's official site: how the product presents itself. hebbrix.com, captured Oct 1, 2026

What you provide and what you get back

What you give Hebbrix and what comes back, from its documentation

You provide

Memories and documents sent via API or SDK
Chat messages through the OpenAI-compatible endpoint
Reported outcomes against decisions

You get back

Ranked recalled memories injected into context
Knowledge graph of entities and relationships
User profiles
Exports of all memory

Where you can use it

Platforms: Web, API
Account: Required
API: OpenAI-compatible chat completions plus memory REST API
SDKs: Python and Node shown
MCP: Hosted server with 33 tools
Hosting: AWS in the United States
Retrieval signals: Semantic, lexical, graph, temporal, outcome

What people use Hebbrix for

Scenarios are worked out by our editors from the documentation; the actual result depends on your input and plan.

Give production agents persistent memory across sessions

Who: Developers shipping AI agents
You provide: Conversation turns and facts stored per collection
Expected result: Agent recalls prior context without extra calls
Plan needed: Starter or Pro
Limitation: Needs API integration work source (opens in a new tab)

Customer support assistants that recall past cases

Who: Support automation teams
You provide: Past tickets and resolutions as memories
Expected result: Replies informed by what resolved earlier cases
Plan needed: Pro
Limitation: Outcome Memory API starts at Pro source (opens in a new tab)

Voice agents with long-term user context

Who: Voice agent builders
You provide: User preferences and call history
Expected result: Personalized responses over repeated calls
Plan needed: Starter
Limitation: Latency depends on your stack; see docs on latency and SLA source (opens in a new tab)

Multi-tenant agent platforms

Who: Platform teams
You provide: Separate collections per customer
Expected result: Isolated memory per tenant
Plan needed: Scale
Limitation: Over 1M credits requires talking to the vendor source (opens in a new tab)

Key features and their practical limits

Feature overview on the official site. hebbrix.com, captured Oct 1, 2026

OpenAI-compatible drop-in API (change base URL)

Routes your existing OpenAI client through the Hebbrix endpoint so relevant memory is injected into context automatically. Adds memory without rewriting the agent.

Access: All plansLimit: Only covers chat-completions style calls source (opens in a new tab)

Outcome-weighted hybrid recall

Ranks recall with semantic, lexical, graph, temporal and reported-outcome signals. Surfaces what worked rather than only similar text.

Access: Hybrid search on all plans; Outcome Memory API from ProLimit: Outcomes must be reported by your agent; missing feedback is not treated as failure source (opens in a new tab)

Knowledge graph of entities and relationships

Extracts entities and relationships and tracks when facts are replaced. Lets agents reason over who knows whom and what changed.

Access: Starter and aboveLimit: Manual graph mutation has additional gates source (opens in a new tab)

Memory decay and user profile extraction

Ranks memory by usage, recency, importance and corrections, and builds user profiles. Keeps stored memory relevant over time.

Access: Profile extraction on all plansLimit: Decay affects ranking, not silent deletion source (opens in a new tab)

Hosted MCP server

Exposes memory to MCP-compatible agents through one endpoint with 33 listed tools. Connects coding agents like Claude or Cline.

Access: Documented in docsLimit: Plan availability not detailed on the fetched page source (opens in a new tab)

Hebbrix pros and cons

Reasons to choose Hebbrix

Free plan with 1,000 credits per month, no credit card stated You can prototype before paying. documented source (opens in a new tab)
Memory export available on every plan Stored memory and graph stay portable if you leave. documented source (opens in a new tab)
Bring-your-own model key on supported paid plans You control which LLM and its costs. documented source (opens in a new tab)
Public docs with API reference Integration can be evaluated before signup. documented source (opens in a new tab)
Customer Content is not used to train shared models Reduces data-use risk for production data. documented source (opens in a new tab)

Check before paying for Hebbrix

Usage is metered in credits with documented weights Cost needs estimating from operation weights. documented source (opens in a new tab)
On Free and Starter, memory writes pause when quota is exceeded Agents stop learning until the next cycle. documented source (opens in a new tab)
Outcome Memory API listed from the Pro plan The headline feature costs $99/month or more. documented source (opens in a new tab)
Aimed at developers; requires API integration Not usable by non-technical teams directly. documented source (opens in a new tab)
Data hosted on AWS in the United States Data residency outside the US is not offered on fetched pages. documented source (opens in a new tab)

How to use Hebbrix for adding persistent memory to an AI agent

Hebbrix documentation: where the steps below come from. hebbrix.com, captured Oct 1, 2026

Set up Hebbrix

Create a Hebbrix account on the Free plan and generate an API key in the dashboard.
Install the Python SDK (MemoryClient) or keep your OpenAI client.
Create a collection for your agent via client.collections.create.
Set base_url to https://api.hebbrix.com/v1 if you use the OpenAI client.

Example: adding persistent memory to an AI agent

Store a memory such as 'User prefers dark mode' in the collection.
Search with the query 'user preferences' in the same collection.
Send a chat completion and confirm the stored fact is injected into context.
Report an outcome against the decision so recall can weigh it.

Review and export the result

Check the dashboard for quota and credit usage after the calls.
Export memories to confirm the stored items are present.

Hebbrix integrations and API

Hebbrix documentation (continued). hebbrix.com, captured Oct 1, 2026

Native integrations and export formats

Export: Memory and knowledge graph export (format not stated)

API access and separate billing

Hebbrix has no public API in its official documentation. OpenAI-compatible chat API plus memory REST API API documentation ↗

Privacy, data use and commercial rights

Summarized from the vendor's current policies; this is not legal advice and a vendor's badge is not an independent audit.

How your inputs and outputs are handled

Stores memories, prompts, outputs, documents and embeddings you provide.

Customer Content is not used to train shared or general-purpose models.

Retained while account is active; account deletion starts a 90-day recoverable window; app logs up to 90 days.

Commercial use and plan restrictions

Not restricted on fetched pages.

Controls for business users

Export and deletion controls
SSO on Enterprise

Policies: hebbrix.com/privacy

Best Hebbrix alternatives for different needs

"Better suited" is based on documented features and pricing, not on a head-to-head test.

ChatGPT: better suited to

Claude: better suited to

Alternatives
Alternative	When to choose it	Pricing	Limitation
ChatGPT		—	—
Claude		—	—

Explore Hebbrix further

Hebbrix pricing, plan by planEvery plan with USD prices, limits and what is free.Best Hebbrix alternativesTools for the same job, by need and starting price.Best AI coding toolsOur picks in this category, compared by price and free tiers.

If you are looking at Hebbrix

More coding tools

Guides and comparisons

AGENTS.md basics

Hebbrix user reviews

No user reviews yet.

Used Hebbrix? Be the first to share your experience.

Write a review

Hebbrix FAQ

What counts as a credit in Hebbrix?

Credits are normalized usage units with documented weights per operation; managed model calls also track tokens. The dashboard shows quota and usage. 1,000 credits is roughly 300 chat turns with memory. source (opens in a new tab)

What happens when I exceed my Hebbrix quota?

On Free and Starter memory writes pause until the next cycle while reads keep working. On Pro and Scale you can enable overage billing so nothing is interrupted. source (opens in a new tab)

Can I use my own LLM with Hebbrix?

Yes, bring-your-own-key works on supported paid plans and Enterprise. Memory is model-agnostic, so stored data stays portable across models. source (opens in a new tab)

Can I export my memory from Hebbrix?

Yes, every memory and the knowledge graph can be exported at any time on any plan. source (opens in a new tab)

How do I connect an existing OpenAI app to Hebbrix?

Change the client's base_url to https://api.hebbrix.com/v1 and use your Hebbrix key; memory is then retrieved and injected automatically. source (opens in a new tab)

Sources and review methodology

Official sources and verification dates

https://hebbrix.com homepage · checked Oct 1, 2026
https://hebbrix.com/pricing pricing · checked Oct 1, 2026
https://hebbrix.com/docs docs · checked Oct 1, 2026
https://hebbrix.com/privacy policy · checked Oct 1, 2026

Not verified: Credit weights per operation; US-specific availability; Export file format; Changelog.

How we documented and maintain this page

Every fact about Hebbrix above comes from the 4 official pages listed here, last compared on Sep 29, 2026; we have not run Hebbrix ourselves, so it has no editorial score. How we check and re-check pages.

Article updated Oct 10, 2026; prices checked Sep 29, 2026. Outbound links to vendors are plain links; we do not earn commissions. How we review · Report an error