Quick Verdict
Hyperspell is an active proprietary company-memory and context service operated by Digital Workers of California, Inc. As of July 21, 2026, it connects business systems such as Slack, Gmail, Google Drive, Notion, Linear, Jira, Salesforce, HubSpot, and GitHub, synthesizes their information into a permission-aware context graph, and serves results to agents through an API, Python and TypeScript SDKs, MCP, and a filesystem-oriented interface.
The licensing boundary matters. Some public client SDKs are MIT-licensed, but the hosted platform, connectors, indexing, storage, and context-graph service are proprietary. The SDK license does not make the whole service open source or self-hostable. Hyperspell publishes no standard plan prices on its main site; the buying path is to book a demo or contact sales. It is therefore an enterprise vendor-evaluation decision, not a free developer utility.
Best For
Hyperspell best fits platform teams building internal agents for support, sales, engineering, or operations when knowledge is scattered across collaboration, CRM, document, and development systems. It may reduce the burden of maintaining connectors, indexing, freshness, and permissions internally. Developers who need a narrower memory component should compare Mem0. Teams that want a broader AI application platform can assess Dify, while n8n is a stronger fit for visible, deterministic integration workflows. Regulated organizations need a legal and security review before connecting sensitive records.
Key Features
- Business-system connectors: Hyperspell claims more than 50 prebuilt connectors spanning email, collaboration, documents, CRM, and engineering systems.
- Permission-aware context: Source permissions are inherited while people, projects, decisions, and facts are synthesized into a connected model.
- Persistent agent memory: The service stores source summaries and extracted memories for future queries and conversations.
- Agent-ready output: It returns structured results or LLM-ready Markdown for custom agents and internal applications.
- Developer interfaces: Universal API, MCP, and MIT-licensed Python and TypeScript SDKs support multiple agent frameworks.
- Freshness and provenance claims: Real-time updates, inherited permissions, and traceable facts are product claims that should be measured during a pilot.
Use Cases
An internal support agent can retrieve past tickets, product documents, and messages while respecting employee permissions. A sales assistant can combine CRM history with meeting notes and email. Engineering agents can retrieve context across GitHub, Linear, Jira, and design documents. A sensible pilot should test permission revocation, connector lag, source links, deletion behavior, and retrieval quality with low-sensitivity data before expanding access.
Pricing
Hyperspell did not publish seat, storage, request, or API pricing on its main website at this review’s cutoff. Pricing is contact-sales, with a demo as the primary entry point. Buyers should request written details for connector limits, indexed volume, query usage, test environments, support, SLA, overages, exports, and termination assistance.
The MIT SDKs do not remove hosted-service charges. If self-hosting or dedicated infrastructure is required, obtain an explicit answer from the vendor rather than inferring deployment rights from a client-library license.
Pros
- Combines connectors, continuous indexing, permission inheritance, and agent context in one managed layer.
- API, MCP, and Python/TypeScript SDK support multiple application stacks.
- The company states that it stores summaries and memories, giving buyers a more concrete data model to evaluate.
- The privacy policy says customer data is not used to train large language models and provides deletion-request channels.
Cons
- The core hosted platform is proprietary even though selected SDKs use MIT licenses.
- Pricing and limits are not public, making cost and exit planning sales-dependent.
- Public terms state that the service is hosted in the United States and do not advertise selectable residency regions.
- The privacy policy names AWS, Pinecone, OpenAI, Sentry, and Langfuse among subprocessors or service providers, creating a multi-party data path.
- Terms say the service is not tailored for HIPAA or FISMA and may not be used in a way that violates GLBA.
Hyperspell’s website claims SOC 2 certification and GDPR compliance, but these remain vendor claims until a buyer reviews the audit scope, report, exceptions, DPA, and technical controls. Its policy says information is generally retained while an account exists or as needed, with backup data potentially isolated until deletion is possible. Buyers should confirm raw-source caching, logs, backups, deletion SLAs, subprocessor changes, encryption, and breach notification contractually.
Alternatives
| Alternative | Better for | Main difference from Hyperspell |
|---|---|---|
| Mem0 | Adding long-term user memory to an application | Narrower memory layer; enterprise connectors and a permission graph are less central |
| Dify | Full AI applications, knowledge bases, and workflows | Broader platform with some self-hosting options |
| n8n | Cross-system automation and visible workflows | More explicit data movement, less focused on a persistent company-memory model |
| Notion AI | Teams whose knowledge already lives in Notion | Simpler scope, with less independent cross-source context infrastructure |
FAQ
Is Hyperspell open source?
Not as a complete product. The managed platform is proprietary; selected Python and TypeScript SDKs are MIT-licensed.
Does Hyperspell offer a public free plan?
No public standard plan or free allowance was listed at the July 21, 2026 cutoff. The site directs prospects to a demo and sales contact.
Does Hyperspell store company data?
Yes. Its FAQ says it stores summaries of source data and extracted memories. Raw caches, logs, backups, and retention should be clarified in the contract.
Is customer data used to train large models?
The vendor’s privacy policy says data will not be used to train large language models. Enterprises should ensure that commitment also covers subprocessors in the DPA.
Can it process healthcare, government, or regulated financial data?
Do not assume so. The terms expressly state that the service is not tailored for HIPAA or FISMA and prohibit use that would violate GLBA.
Can customers choose a data-residency region?
Public terms identify US hosting but do not list selectable regions. Ask the vendor for a written architecture and residency commitment.
Bottom Line
Hyperspell offers a clear proposition: managed, permission-aware company memory for agents across many business systems. It can replace substantial connector, indexing, and authorization work, but it places valuable organizational context in a proprietary US-hosted service and its vendor chain. Start with low-sensitivity data, contractually verify storage and deletion behavior, test permission revocation and freshness, and expand only after security, legal, and cost reviews. Its published HIPAA, FISMA, and GLBA limitations should be treated as hard procurement constraints.