Pastille

What Pastille does

Pastille is an early-stage private company building a bi-temporal, reified, compositional knowledge graph that functions as a typed, self-maintaining company context layer for autonomous AI agents. The platform is accessed by agents via the Model Context Protocol (MCP), allowing integration without requiring users to learn a new interface. Its core product proposition is enabling AI agents to make informed decisions by retrieving structured, queryable company knowledge rather than operating on shallow or unstructured context.

The underlying architecture combines bi-temporality (tracking both valid and transaction time), reification (treating statements as first-class entities), and composition, which together support a rich representation of evolving enterprise knowledge. Pastille targets enterprise organizations deploying AI agents as its primary customer segment, with a horizontal segmentation approach focused on the use case of agent-driven decision-making. The platform's text-based knowledge retrieval capability is exposed through MCP, positioning it as interoperability infrastructure within the broader agent tooling ecosystem.

The business model is not publicly disclosed — pricing is unavailable and there is no evidence of named customers, headcount, or revenue. The company has raised approximately $194,114 from Antler in a round dated 2025-12-22, consistent with a pre-seed or early-seed stage. The minimal public footprint and absence of customer logos or case studies indicate Pastille is at a very early stage of commercialization, with its strategic positioning centered on becoming a standard context layer for the emerging autonomous agent market.

Name

Pastille

Website

https://pastille.ai

Company type

Private

Operating status

Operating

Short description

Pastille provides a bi-temporal, reified, compositional knowledge graph that serves as a self-maintaining company context layer for autonomous AI agents, accessible via MCP. The company targets enterprise organizations deploying AI agents.

Ownership category

akta.pro rank

Product category

AI Knowledge Management Software

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Web Search Portals and All Other Information Services (519290)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

Enterprise AI Data & Knowledge Platforms (Vector Databases, Knowledge Graphs) (HDAEANAH)

Keywords

• Knowledge graph platform
• AI agent context layer
• Enterprise knowledge management
• Model Context Protocol
• Semantic AI infrastructure

Headquarters

HQ city

Sweden

HQ country

United States

HQ region

North America

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Operations, Marketing or Sales, Infrastructure

Core offering

Pastille provides a bi-temporal, reified, compositional knowledge graph that delivers a self-maintaining enterprise context layer to autonomous AI agents via the Model Context Protocol (MCP). It functions as a typed company brain that AI agents can query to understand company-specific information and make informed decisions, without requiring users to learn a new interface.

Product overview

Pastille is a unified platform built on a bi-temporal, reified, compositional knowledge graph that serves as a typed, self-maintained company brain for AI agents. The core product (Pastille) integrates with any agent via MCP (Model Context Protocol), requiring no new interface to learn, and provides structured knowledge retrieval capabilities.

Differentiator

Problem solved

Functional benefit

Products and services

• Pastille A knowledge graph platform built on a bi-temporal, reified, compositional architecture that provides AI agents with structured company knowledge via MCP, enabling the agents to understand and make decisions about company-specific information. It is targeted at enterprises deploying autonomous AI agents.

Segments1 record

Ideal customer profiles1 record

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration1 record

AI capability2 records

Feature3 records

Recent moves3 records

Expansion highlights3 records

Direct peers

• Mem0: Mem0 builds a self-improving memory layer for AI agents and assistants, directly competing with Pastille as an enterprise context/memory layer delivered to agents for personalization and decision-making.
• Zep AI: Zep provides long-term memory and a context server for AI assistants and agents, competing head-to-head with Pastille's positioned 'company brain' for production agents.
• Cognee: Cognee is an open-source knowledge/memory layer that builds a graph-based 'cognitive architecture' for LLMs and agents, overlapping directly with Pastille's graph-based company context approach.

Emerging players

• Letta: Letta (formerly MemGPT) provides stateful agent infrastructure with persistent memory; competes with Pastille on the memory/persistence leg of the agent context problem.
• LlamaIndex: LlamaIndex provides a data framework and ingestion/retrieval toolchain for LLM applications including knowledge graphs and agent context, overlapping with Pastille's structured company knowledge value proposition.
• Kuzu: Kuzu is an embedded graph database increasingly used inside LLM and agent stacks for structured knowledge; comparable to Pastille's graph-based architecture serving agent context.

Broad incumbents

• Pinecone: Pinecone is a managed vector database widely used as the retrieval layer under AI agents and RAG systems; represents the incumbent retrieval stack that Pastille's typed-graph approach must displace or sit beside.
• Neo4j: Neo4j is a leading graph database platform with growing focus on knowledge graphs for GenAI; directly overlaps Pastille's graph-based company brain and brings enterprise scale, distribution, and tooling.
• Weaviate: Weaviate is an open-source vector + hybrid (vector + symbolic/graph) database targeting AI-native applications; comparable to Pastille as a retrieval/knowledge substrate for agents.
• LangChain: LangChain/LangGraph is the dominant open agent framework and offers memory, retrieval, and context-management primitives that intersect with Pastille's enterprise context layer for agents.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat3 records

Key risks6 records

Key highlights5 records

Customer concentration

Revenue estimate

Valuation estimate

Number of profiles

Funding overview

Funding rounds1 record

Investors1 record

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Pastille