Skip to main content

Hands-on demos of Squirro Neo, GraphRAG and our agents. No sign-up, no sales call. Try Them Now

Run Every Query and Agent on One Foundation You Govern.

Squirro is the AI context platform beneath your queries and agents, providing a governed layer that brings in your content, applies permissions, and delivers trusted, sourced context wherever your data stays.

How Squirro turns your data and knowledge models into one AI context layer Your data, from data platforms such as Snowflake, Databricks and Palantir, documents and content in SharePoint, Confluence, Jira and Exchange, and live systems such as SAP and Salesforce, is ingested through connectors or queried live by agents over MCP servers and REST APIs. Ingestion loads, enriches, classifies, chunks and embeds every item, with access rights and labels on every chunk. The AI Context Layer holds one unified index of content, access rights, vectors and entities, a knowledge graph for GraphRAG that uses your taxonomy, ontology and existing knowledge graphs, and vector search with hybrid search and re-ranking. Squirro Solutions, Enterprise AI Search, AI Agents, Agentic Workflows and Dashboards, build on it with cited answers, guardrails and human approval. Squirro sends context to the models you choose: self-hosted open-weight models inside your perimeter, or external LLMs. Through the Squirro API and MCP it reaches people in Squirro Neo, AI coding assistants, and enterprise apps and agents. The platform runs in your VPC or on premises. SQUIRRO PLATFORM AI CONTEXT LAYER ingest query live context KNOWLEDGE MODELS Taxonomy and ontology Existing knowledge graphs YOUR DATA Ingested into the index, or queried live by agents Data platforms Snowflake Databricks Palantir Documents and content SharePoint Confluence Jira Exchange Live systems SAP Salesforce MCP servers and REST APIs CONNECTORS Ingestion Load Enrich, OCR Classify Chunk, embed Access rights and labels on every chunk Unified index Content Access rights Vectors Entities One index across every source Knowledge graph · GraphRAG Your ontology Imported graphs Traverses linked entities and relations Vector search · RAG Hybrid search Re-ranking Finds semantically similar passages Squirro Solutions Enterprise AI Search AI Agents Agentic Workflows Dashboards Cited answers, guardrails, and human approval built in VPC or on premises Your models Squirro sends context, the model reasons. SELF-HOSTED MODELS Open-weight models, inside your perimeter EXTERNAL LLMS Anthropic OpenAI and others SQUIRRO API · MCP PEOPLESquirro Neo Chat, search, and dashboards for every employee Single sign-on, directory permissions AI CODING ASSISTANTS Build on Squirro Claude Code, Codex, Copilot, and others get governed access to build agents and dashboards. ENTERPRISE APPS AND AGENTS Every app and agent, grounded One governed context for business apps, add-ins, and agents across the enterprise.
Your data goes into one AI context layer, and comes out as answers and agents Your data, from systems such as SharePoint, Confluence, Jira, SAP, Salesforce, Snowflake, Databricks and Exchange, flows into the Squirro Platform, which runs in your VPC or on premises. Its AI Context Layer holds a unified index, a knowledge graph and vector search. Squirro Solutions build on it and reach people in Squirro Neo, AI coding assistants, and enterprise apps and agents. SQUIRRO PLATFORM AI CONTEXT LAYER YOUR DATA SharePoint Confluence Jira SAP Salesforce Snowflake Databricks Exchange and more VPC or on premises Unified index Knowledge graph · GraphRAG Vector search · RAG Squirro Solutions Enterprise AI Search AI Agents Agentic Workflows Dashboards PEOPLE Squirro Neo AI CODING ASSISTANTS Build on Squirro ENTERPRISE APPS AND AGENTS Every app and agent, grounded

See the Full Architecture

Trusted by Industry Leaders

library of congress OCBC Bank Zwickroell Security Benefit Henkel Siemens Bertelsmann pavlov media long beach container terminal Deutsche Bundesbank Candriam IDB Invest Bühler logo swissrockets Customer Logo

Explore Each Part of the Platform.

Connectors

Connect to all your data sources.

Squirro connects to the systems your data already lives in, including content repositories, CRMs, ticketing, databases, BI, and market data, and brings it all in, ready for retrieval.

  • SharePoint
  • Confluence
  • Salesforce
  • Exchange
  • Google Drive
  • Documentum
  • Jira
  • Zendesk
  • Amazon S3
  • SAP
  • ServiceNow
  • Dropbox
AI Context Layer

One index, your knowledge graph, and vector search behind every answer.

Every source lands in one index, with its access rights. Your taxonomy and knowledge graph, managed in Squirro Graphite, add how your business is organized, so an answer follows the links between things and cites the passage it came from.

Squirro Graphite, the knowledge graph manager: the Concept Visualizer shows one person’s links to Sales, Pricing Management, Air Transportation, and Pittsburgh
Agents and Models

Run every agent on the models you choose.

Squirro is LLM agnostic and hosts no models. Your administrators choose the models each project can use, self-hosted inside your perimeter or external, and pin each agent to the one that fits its job.

Security and Trust Center

Permissions on every query, for people and agents.

Every search is filtered before it reaches the index, by the access each project grants and by source permissions where the connector carries them. Agents search with the same filter as people, and a person approves the actions you choose.

  • ISO 27001 logoISO 27001Certified
  • SOC 2 logoSOC 2 Type IAttested, Type II in progress
  • GDPR logoGDPR & CCPACompliant
  • Encryption shield and padlockEncryptionTLS 1.2+ in transit, AES‑256 at rest
Deployment Options

The complete platform, wherever your data has to stay.

Squirro Cloud, your VPC, or your own premises. Each runs the complete platform, not a reduced version, so data sovereignty never costs you features.

Squirro Neo

Where your people work.

Search, chat, tasks, and agents in one place, on the same permissions and project settings as everything else on the platform.

Squirro Neo answering a process engineer: the root cause of a recurring fault on Line 3, the steps it took in the plant systems, and the answer with numbered sources

Build Your Own on the Same Foundation.

Agents as configuration

Describe it, then deploy it.

Agents, models, and tools are declared in one project configuration, checked and deployed in one pass.

genai lint passes on a new agent: project.yaml is ok, the agent file is ok with its tool set and one approval rule, and there are no errors or warnings
Explore Agents
Your dashboards

Built around each team.

Your developers build them with the coding assistant they already use, and they open in Squirro beside chat and search. Developer preview.

Client Book, a screen a bank’s developers built, open in Squirro: clients, assets under management, rate sensitivity, and covenant alerts, then each client’s euro exposure and next meeting
Everything else

The Squirro APIs.

Every interface reaches the platform through the same APIs, so your applications get the same search, permissions, and agents.

Read the Developer Docs

The Full Architecture, From Source to Answer.

How ingestion, the unified index, the knowledge graph, vector search, and your models work together, and where each one runs.

How Squirro turns your data and knowledge models into one AI context layer Your data, from data platforms such as Snowflake, Databricks and Palantir, documents and content in SharePoint, Confluence, Jira and Exchange, and live systems such as SAP and Salesforce, is ingested through connectors or queried live by agents over MCP servers and REST APIs. Ingestion loads, enriches, classifies, chunks and embeds every item, with access rights and labels on every chunk. The AI Context Layer holds one unified index of content, access rights, vectors and entities, a knowledge graph for GraphRAG that uses your taxonomy, ontology and existing knowledge graphs, and vector search with hybrid search and re-ranking. Squirro Solutions, Enterprise AI Search, AI Agents, Agentic Workflows and Dashboards, build on it with cited answers, guardrails and human approval. Squirro sends context to the models you choose: self-hosted open-weight models inside your perimeter, or external LLMs. Through the Squirro API and MCP it reaches people in Squirro Neo, AI coding assistants, and enterprise apps and agents. The platform runs in your VPC or on premises. SQUIRRO PLATFORM AI CONTEXT LAYER ingest query live context KNOWLEDGE MODELS Taxonomy and ontology Existing knowledge graphs YOUR DATA Ingested into the index, or queried live by agents Data platforms Snowflake Databricks Palantir Documents and content SharePoint Confluence Jira Exchange Live systems SAP Salesforce MCP servers and REST APIs CONNECTORS Ingestion Load Enrich, OCR Classify Chunk, embed Access rights and labels on every chunk Unified index Content Access rights Vectors Entities One index across every source Knowledge graph · GraphRAG Your ontology Imported graphs Traverses linked entities and relations Vector search · RAG Hybrid search Re-ranking Finds semantically similar passages Squirro Solutions Enterprise AI Search AI Agents Agentic Workflows Dashboards Cited answers, guardrails, and human approval built in VPC or on premises Your models Squirro sends context, the model reasons. SELF-HOSTED MODELS Open-weight models, inside your perimeter EXTERNAL LLMS Anthropic OpenAI and others SQUIRRO API · MCP PEOPLESquirro Neo Chat, search, and dashboards for every employee Single sign-on, directory permissions AI CODING ASSISTANTS Build on Squirro Claude Code, Codex, Copilot, and others get governed access to build agents and dashboards. ENTERPRISE APPS AND AGENTS Every app and agent, grounded One governed context for business apps, add-ins, and agents across the enterprise.

Recognition

Cited in 26 Gartner® Reports in 2026.

Across horizontal AI research and industry reports, from banking and manufacturing to oil and gas and government.

26 Gartner® research notes citing Squirro in 2026
15 Hype Cycles™ citing Squirro, from Generative AI to Banking
Top 5 most-cited knowledge management vendors in Gartner client inquiry
Named for
  • Enterprise AI Platform
  • AI Context Platform
  • Retrieval-Augmented Generation
  • GraphRAG
  • Knowledge Graphs
  • AI Knowledge Management

Frequently Asked Questions.

Is there one platform under search, chat, and agents?

Yes. Search, chat, agents, dashboards, and your own applications all reach the platform through the same APIs, so the same permissions and project settings apply to all of them.

Do we have to move data to add new agents?

No. New agents work on your existing projects, content, index, and permissions.

How do source permissions work?

Project access is granted to users and groups, and every query is filtered before it reaches the index. Where a connector carries the source system’s permissions, for example document-level access in a document store, those apply as well. Your engineer will confirm this per source.

Does Squirro host the language models?

No. Squirro is LLM agnostic: your administrators choose the models each project can use and pin each agent to the one that best fits its job, and the agents call those endpoints. Squirro does not pick a model on its own.