Evokoa
Evokoa is a research and engineering company building infrastructure that makes enterprise data traversable and useful to AI agents. Its stated long-term mission is to make the world's data AI-native by adding context and relationship-aware retrieval without requiring teams to move their source data.
At a Glance
- AI infrastructure and agent builders
- Founders and developers building on Postgres
- Financial services, fintech, marketplaces, and payment products
- Support, customer-success, and operations teams
- +3 more
AI Tools by Evokoa
(2)
WonderSearch
Serverless Semantic Document Search
Polygres
Hybrid Vector Graph Search API
Discussions
No discussions yet
Be the first to start a discussion about Evokoa
Latest News
WonderSearch published 'Small, Medium, or Large: choosing your search effort,' documenting selectable per-query search effort and usage-based billing.
WonderSearch published 'Give your agent your documents, without giving it everything,' describing MCP-based access to approved knowledge collections.
WonderSearch published 'Write your evaluation set before you pick a search tool,' recommending evaluation questions before selecting a search product.
WonderSearch published 'You wanted search. How did it become an embedding project?', positioning its semantic search as an alternative to building an embedding pipeline.
Products & Services
Serverless semantic search for files and data. Users ask questions in natural language and receive ranked passages with document references, without building an embedding pipeline or managing search servers. It is available through MCP, an API, a CLI, and a Python SDK.
Evokoa's platform for retrieval over Postgres. It combines AI Context Search for semantic, text, and grouped retrieval with graph retrieval over modeled relationships; it supports hosted Postgres projects and synchronized projects connected to supported Postgres sources.
An open-source, Rust-based graph traversal extension/engine for PostgreSQL. It builds a compact derived relationship index over relational tables and supports capabilities such as bounded BFS/DFS, shortest paths, and connected components while keeping Postgres as the source of truth.
The semantic-retrieval engine and collection-based search path used in the Polygres platform, combining meaning, text signals, grouped results, and graph context to return inspectable source records and passages.
Market Position
Evokoa positions its products as a lighter alternative to vector-database/RAG pipelines and traditional graph databases: WonderSearch provides semantic search without embeddings, ingestion pipelines, or search servers, while Polygres and pgGraph combine retrieval and graph traversal directly around Postgres. Evokoa specifically contrasts pgGraph's in-memory relationship layer and deep traversal model with Apache AGE's recursive execution inside Postgres, and emphasizes keeping Postgres as the source of truth rather than requiring a separate heavy graph database.
Leadership
Founders
Dale Everett Ng
CEO and founder. He grew Swig Energy from launch to 5,000+ customers, previously led community at Bholdus, bootstrapped logistics and marketing businesses, and co-founded Master Technologies (AI) at 17. He received the Toh Chin Chye and IMDA Gold Medals at Singapore Polytechnic.
Dalton Prescott Ng
CTO and founder. A former researcher at DSO, he started coding at 11 and founded his first company at 14. He was an Apple Scholar in 2018 and 2019 and previously served as Head of Product and Pre-Sales at BitCyber, managing more than 250 B2B partners across APAC.
Executive Team
Dale Everett Ng
CEO
Previously led community at Bholdus, grew Swig Energy to 5,000+ customers, and bootstrapped logistics and marketing businesses.
Dalton Prescott Ng
CTO
Former DSO researcher and former Head of Product and Pre-Sales at BitCyber; coding since age 11 and an Apple Scholar in 2018 and 2019.
Board of Directors
Founding Story
The founders say Evokoa grew out of work on operational intelligence for multi-location enterprises. They found that multi-hop relationship questions were too slow for AI-agent workloads and that existing graph products required moving the system of record into a separate database, so they built a lightweight relationship layer and graph traversal engine over Postgres. Their broader vision is to make existing business data traversable by agents without ETL, a second source of truth, or a long integration project.
Business Model
Revenue Model
WonderSearch charges for search compute per query and bills storage separately; Polygres offers hosted/synchronized retrieval projects and usage-based platform services, with enterprise plans handled through the sales team.
Pricing Tiers
Usage is priced per query, with Small, Medium, and Large search-effort choices.
Storage is billed separately from search compute.
Enterprise customers are directed to contact Evokoa for a plan based on scale and requirements.
Target Markets
- AI infrastructure and agent builders
- Founders and developers building on Postgres
- Financial services, fintech, marketplaces, and payment products
- Support, customer-success, and operations teams
- SaaS companies with complex account and permission models
- Developer platforms, logistics systems, and infrastructure teams
- Agentic memory and AI infrastructure
- AI product workflows built on Postgres
- Fraud detection and connected-risk analysis in financial services and marketplaces
- Support, success, and operations copilots
- Permission-aware applications and multi-hop access checks
- Dependency analysis, blast-radius analysis, and infrastructure recovery workflows