Tessera DB
PX-01
Documentation
The reference, behind a passphrase.
The Tessera DB reference documentation is access-controlled. Enter the passphrase if you have one, or request access by email at ceo@par2labs.com. The tutorials and writing below are open to everyone.
Reference documentation · access-controlled
The full technical reference, for teams evaluating or running Tessera DB.
It is encrypted at rest and opened in your browser with a passphrase — nothing is sent anywhere to check it, and it is kept only for this visit. The tutorials and writing below are open to everyone.
AES-256-GCM
PBKDF2 · 600k
decrypted in your browser
Tutorials and writing.
Tutorials
15
Tessera DB Tutorial: Build Your First Hyperedge in 5 Minutes
Start the binary, load a sample dataset from the admin dashboard, and watch the same records appear across the relational, graph, vector, full-text, and time-series dimensions — one copy of the data, five ways to query it.
Beginner · 7 min
Tessera DB Tutorial: Connect a Postgres Database with Automatic Graph Endpoint Mapping
Wire up a Postgres source so Tessera mirrors your tables, auto-derives graph relationships from your foreign keys, and keeps in sync with incremental polling.
Beginner · 7 min
Tessera DB Tutorial: Bitemporal Queries to Reconstruct Historical State for Audit
Every fact carries valid time and transaction time. Learn to answer 'what was true on date X' and 'what did the system show on date X' — two different questions, two different queries.
Intermediate · 5 min
Tessera DB Tutorial: Build a GraphRAG Retrieval System End to End
Ingest documents, auto-embed them, run hybrid vector + keyword + graph retrieval fused with Reciprocal Rank Fusion, and assemble LLM context with provenance — all in one engine.
Intermediate · 6 min
Tessera DB Tutorial: Connect a REST API Source with Bearer Auth
Pull rows from any JSON-returning endpoint, with bearer / basic / header auth and a configurable rows path for nested response shapes.
Beginner · 3 min
Tessera DB Tutorial: Set Up Auto-Embedding for Semantic Search
Configure an embedding provider once. Every record with text properties gets indexed for vector search automatically — no separate pipeline.
Beginner · 2 min
Tessera DB Tutorial: Discover Schema and Auto-Derive Graph Endpoints from Foreign Keys
Let Tessera read your Postgres schema and turn every foreign key into a real graph relationship — no manual mapping.
Beginner · 2 min
Tessera DB Tutorial: Set Up Incremental Sync with a Cursor Column
Stop re-ingesting your whole source on every poll. A cursor column tracks where the last sync left off and pulls only newer rows.
Beginner · 3 min
Tessera DB Tutorial: Monitor Sync Attempts Across All Sources
Every sync attempt — success or failure — is recorded with its timing and cursor. Use the cross-source feed to see your whole fleet at once.
Beginner · 2 min
Tessera DB Tutorial: Enable At-Rest Encryption for Connector Credentials
Turn on opt-in at-rest encryption and connector passwords, API keys, and stored data become authenticated ciphertext on disk.
Beginner · 3 min
Tessera DB Tutorial: Use the SQL Interface with JOINs and Aggregates
Tessera speaks real SQL. The same JOIN, GROUP BY, and ORDER BY you already know — over the same data as the graph, no new language required.
Beginner · 4 min
Tessera DB Tutorial: Run PageRank over Your Graph in 5 Minutes
Score every entity by its centrality in the graph. Useful for fraud rings, influence networks, and feature engineering for ML.
Beginner · 3 min
Tessera DB Tutorial: Detect Communities in Your Graph with Louvain
Find tightly connected clusters of entities — for fraud rings, customer segments, document topics, anything where 'who hangs out together' matters.
Intermediate · 3 min
Tessera DB Tutorial: Bucket and Aggregate Time-Series Data with Zero Copies
Every fact already carries valid time and transaction time, so any set of records is also an ordered series — bucketed and aggregated in place, no separate metrics store.
Intermediate · 3 min
Tessera DB Tutorial: Branch Your Data Like You Branch Code
Lightweight copy-on-write branches let you run experiments, schema changes, or what-if analyses on a fork of your live data without touching the original.
Intermediate · 3 min
Writing
14
Why We Built a Hyperedge-Native Database from Scratch
Property graphs treat every relationship as a binary edge. That model breaks the moment a single fact has three or more participants. Here's what changed when we stopped pretending otherwise.
2 min
Five Databases Are Four Too Many
Postgres for relational, Neo4j for graphs, Pinecone for vectors, Elasticsearch for text, TimescaleDB for time-series. The modern AI stack is a hidden tax on every team that ships it.
3 min
The Case Against Vector Store Sprawl in AI Applications
Pure vector retrieval gives you semantic similarity. It does not give you context. GraphRAG fuses three signals — and it consistently beats vector-only on the queries that matter.
2 min
Bitemporal Data Isn't Optional for Regulated Workloads
Knowing what was true at a point in time, and knowing what your system thought was true at a point in time, are different questions. Regulators ask both. Most databases can answer neither.
2 min
TQL Was Designed for LLMs to Write — Here's How It Works
Every AI agent on the planet is generating queries against your database. The query language they reach for matters more than most teams admit. Here's why we wrote a new one.
2 min
Content-Addressed Identity Ends the Integer-ID Drama
Auto-incrementing IDs feel obvious until you've spent a sprint reconciling them across two systems. Cryptographic hashes solve the problem at the source.
1 min
Why We Kept Cypher Around
If TQL is better for LLMs and SQL is better for analysts, why ship a third query language? Because the world has a decade of Cypher queries already written, and breaking them helps no one.
2 min
The Half-Open Interval Is a Small Choice with Big Consequences
Does a time range include its endpoint? Different databases answer differently. The wrong answer creates phantom rows or missing rows. We picked [start, end).
1 min
Confidence Scores Belong on Every Fact
Half the data in a modern application is inferred by an AI model. Treating those rows the same as a manually-entered bank balance is a bug we keep shipping.
2 min
Schema-on-Write, Schema-on-Read, Schema-on-Demand
Strict types or document flexibility? The right answer is: both, switchable per type, evolvable without downtime. Here's how we got there.
1 min
Reciprocal Rank Fusion: The Unsung Hero of Hybrid Search
Combining vector, keyword, and graph signals into one ranked list isn't hard if you know the trick. RRF is the trick, and it's much simpler than the literature makes it look.
2 min
What We Learned Shipping the First Postgres Connector
Building a database connector sounds straightforward until you actually do it. Cursor semantics, late-arriving data, schema drift — all the things we got wrong on the first pass.
2 min
Defending the Data, Not the Perimeter
Most databases assume the network keeps attackers out and store data in the clear behind it. Tessera's AutoGuard puts an eight-layer AI-security stack inside the engine, so protection travels with the data.
2 min
On-Premise AI Is Not an Oxymoron
The modern AI stack grew up in the cloud. Regulated data can't follow it there. You don't have to choose between the two — the whole hybrid-plus-ML stack fits in a single binary that runs on your hardware.
2 min