kotobase
Menu

STATE WEY YOU FIT TRUST · GRAPH WEY CONTENT DEY ADDRESS

Graph state wey you fit trust for software wey AI write.

Technical details, legal documents and some account steps still dey for English.

Kotobase na graph database wey dey use content take identify data for AI state and knowledge. E keep relationships, where data come from, and history wey you fit identify together. Access dey limited to each tenant, and hosting dey managed.

Database for graph · Datalog · SPARQL · Cypher-style · GraphRAG

You fit test am free. Managed production start from 25 US dollars every month; di Free plan no need card. You go see di real currency dem go charge before you pay.

Code wey safe. State wey dem fit trust. Execution wey dem dey control.

Kotoba Labs dey build safe and ultra-fast stack for AI-generated software. Start with the layer wey your application need.

Kotoba · Safe code

Di language: programs wey you fit check, clear power dem get, plus benchmarks wey you fit do again.Explore Kotoba

Kotobase · Trusted state

Di graph database: clear relationships, content identity, and traceable context.Start free

Kotoba Cloud · Execution wey dem dey control

Identity and deploy control for di execution environment. Discovery dey live; hosted apply no dey offered yet.Explore Kotoba Cloud

Measured on production

8 September 2026 · NRT edge · 100 requests per series · 5 warmups wey dem no count · concurrency 1 · 1 test entity

Biscuit issuance

32.23 ms

p50 · milliseconds

100/100 successful requests

Biscuit verification

18.26 ms

p50 · milliseconds

100/100 successful requests

Authenticated warm query

741.67 ms

p50 · milliseconds

100/100 successful requests

Measured latency — milliseconds
Operation p50 p95 p99 Max
Biscuit issuance 32.23 110.02 123.46 144.59
Biscuit verification 18.26 80.78 105.49 105.64
Authenticated warm query 741.67 961.66 1073.58 3667.31

Single-client production HTTPS latency, including network and authorization. Small two-pattern Datalog point selection; no be throughput, big-graph, controlled cold-cache, or competitor ranking claim.

Every measured query return di inserted marker. All 300 measured requests succeed.

Method plus raw results

AI need more than another chat history

Agents dey forget between sessions. Retrieval dey return disconnected chunks. Operating graph infrastructure dey steal time from di app wey you dey build.

Agents dey forget

Make sure say useful entities, claims, and relationships dey available anytime you need am.

RAG chunks dey disconnect context

Show how people, events, documents, and decisions relate — no be only which text dey look similar.

Graph operations dey get complex

Use hosted secure storage, then managed SLA when production team need am.

Why Kotobase?

Persistent memory

Give agents memory wey go survive di session instead of isolating knowledge for chat history.

Connected context

You fit store and find entities, claims, plus relationships, no be only disconnected pieces.

Agents and apps dey share am together.

Make the same tenant-scoped graph show through APIs and MCP.

Traceable answers

Keep graph state and provenance safe with content-addressed commits.

Vector search dey find similarity. Graphs dey keep relationships intact.

Traditional RAG na good for finding related passages. Graph-backed AI add explicit entities, claims, and relationships so agent fit follow connected context.

Traditional RAG

  1. Documents
  2. Chunks
  3. Vector search
  4. LLM

AI wey get graph support wit Kotobase

  1. Documents · Apps · Data
  2. Knowledge graph
  3. Entities · Claims · Relationships
  4. API · MCP
  5. Apps · AI agents

Kotobase dey complement retrieval; e no talk say every workload suppose replace vector search.

One context layer wey connect everything

Apps · Documents · Data
Kotobase Knowledge Graph
API · MCP
Apps · AI Agents

Tenant-scoped

Applications and agents fit access the graph inside the same tenant identity boundary.

Content-addressed

Signed immutable commit CIDs dey make the selected graph state explicit and reproducible.

Client-contributed

Clients fit query, cache, compute, and contribute verified CID-addressed results.

Join your graph to AI agents through MCP

Give agents graph tool surface without to build one-off retrieval glue for every application.

Your Data
Kotobase Knowledge Graph
API · MCP
Apps · AI Agents

MCP endpoint

https://kotobase.net/mcp

Tools dey generated from the same published lexicons wey the application API dey use.

Make GraphRAG based on clear relationships

Model entities

Represent di people, products, organizations, documents, and events wey your application dey reason about.

Preserve claims

Keep assertions and their graph context dey available across sessions and applications.

Traverse relationships

Use Datalog, SPARQL, or Cypher-style query surfaces for di shape of query wey you need.

Reproduce context

Bind work to content-addressed graph commits instead of implicit mutable database state.

You dey find practical Neo4j alternative?

Kotobase na option to check when you want knowledge graph for agents and applications, MCP access, managed hosting, or content-addressed provenance.

Correct Kotobase evaluation factors wey dem don check
Evaluation factor Kotobase
Start Free secure storage wey dem host; no need credit card at all
Managed production entry $25/month. AuraDB Professional start for $65.70/month (official public price checked 2026-08-27).
Agent access MCP
Knowledge graph Yes
Managed option Yes, e dey — Starter, Professional, Business Critical, Enterprise
Query Datalog, SPARQL, and Cypher style dem dey use
Provenance Content-addressed commit CID dem
Dedicated option Yes — Enterprise contract

Source: Neo4j AuraDB price mata. Prices no include taxes and usage wey pass each plan's included limits.

Migration fit, talk am clear clear

Kotobase no dey presented as wire-compatible with Neo4j. Make you evaluate your data model, query surface, agent integration, and operational requirements before you migrate. Dis page no dey make any unverified price or performance superiority claim.

How di current query protocol dey work plus di limits wey dem don measure

Production scope as of 2026-08-15. Supported syntax, measured evidence, and open gaps dem separate so result wey be only small part no go show as general graph-database performance.

True production capability matrix
Area Current production specification
Storage and truth Signed immutable CID commit DAG; R2-backed immutable blocks and conditional graph heads. Durable Objects, D1, PostgreSQL, and an S3 mutable database no be correctness dependencies.
Di part of Cypher-style wey fit read Labels, parameters, property projection, DISTINCT, ORDER BY/SKIP/LIMIT, NOT/AND/OR predicates, one materializing WITH stage followed by MATCH, simple CASE plus CASE WHEN variable IS NULL, and directed/undirected variable-length paths for primary and OPTIONAL MATCH. OPTIONAL traversal dey run under left join and e dey fill null for misses. For 2026-08-15 production probe, dem combine these new operators and e return the expected row through kotobase.net for graph Worker 838a5a03-f307-46f2-99b8-4717a0573096. This one still na small part, no be full openCypher or performance claim.
Planner and joins Path-free basic graph patterns dem hand over to peer statistics planner as one BGP: e dey measure visibility-scoped clause cardinalities or e dey use fresh scoped materialized statistics, e dey cost-order clauses, then e dey pass estimates to join strategy. Hash natural joins and binding push-down dey used around OPTIONAL/WITH stages. Variable-length path interleaving still dey use bounded heuristics; no be general cost model for every Cypher operator.
Cypher writes Authenticated CREATE of standalone nodes wey need non-empty id dey use the same graph-policy check, immutable transaction/commit blocks, conditional R2 head CAS, nonce replay protection, and conflict retry path like datom transact. For 2026-08-15 production probe, e create created/1 with HTTP 200 and read the graph back. Relationship CREATE, CREATE after MATCH, MERGE, SET, REMOVE, and DELETE dem no allow.
Di part of SQL JOIN wey fit read Bounded SELECT over datoms and request-scoped external tables. INNER/LEFT/RIGHT/FULL/CROSS and non-equality joins, derived tables to depth 2, GROUP BY, and COUNT/SUM/AVG/MIN/MAX dey supported for up to four sources. Datom scans/results stop at 1,000 and intermediate joins at 10,000. External tables na finite scalar rows wey dem supply for the request, no be arbitrary network/database connectors. For 2026-08-14 production probe, e pass FULL external JOIN, grouped derived-table, and LEFT non-equality JOIN shapes; na functional evidence, no be performance claim. HAVING, CTE/UNION, windows, DML, comments, and multi-statements still no support.
Cold reads CID verification, Cache API/R2 block fetch, query-ordered pack Range support, and bounded client caches. The published 12.5M-datom qualification policy need three independent empty-client-cache samples and cold network ceilings of 5 s for point lookup and 12 s for two-hop traversal; e receipt measure cold p99 2.231 s and 6.264 s. This one na reproducible engineering envelope for those two shapes, no be contractual SLA or arbitrary-query guarantee.
Same-graph writes Optimistic conditional-head CAS, full-jitter exponential backoff, 12 attempts by default (operator-bounded 1–16), then HTTP 409 with Retry-After and retry metadata. Writes no be lock-free successes under arbitrary contention; no acknowledged write fit report before e CAS win.
Scale evidence Historical LDBC SF-0.1 receipt: 3,013,602 datoms and five of seven Interactive Short queries. IS2's WITH shape and IS7's OPTIONAL path/CASE IS NULL shape now pass end-to-end tests, but the historical latency receipt never relabel or rerun. SF-1 convert to 25,522,216 datoms; no completed 100M-datom production qualification dey.
Comparison claim No general Neo4j, Neptune, GraphDB, or Datomic performance superiority claim. No audited LDBC Neo4j result dey to compare; published local paired results na only for their exact bounded workloads.

What the benchmark no prove

Warm-cache latency no mean say cold-query claim dey; SF-0.1 no mean SF-1 or 100M scale; five supported LDBC queries no mean full Cypher. Make you run your own data and queries before migration.

Quickstart

1. Start free

Create tenant-scoped hosted graph. No credit card dey required for Free.

2. Connect app or agent

Use API or point MCP-capable agent to https://kotobase.net/mcp.

3. Ingest and query

You fit write entities, claims, plus relationships, then you fit query the same tenant-scoped graph from agents and applications.

Developer details

ipfs pin remote service add kotobase https://kotobase.net <JWT>
ipfs pin remote add --service=kotobase --name=my-doc <cid>
ipfs pin remote ls  --service=kotobase

Authenticate with Passkey or SIWE, then authorize API operations with short-lived tenant- and graph-scoped Biscuit. CACAO and opaque Bearer still na migration-only compatibility paths. Public exact-CID reads dey live at /ipld/:cid; canonical queries select explicit signed commit CID no be mutable database head. Built on kotoba.

Choose di plan wey fit how you dey operate

Start wit Free hosted secure storage. Choose Starter or Professional wen you wan make Kotoba Labs Inc dey operate di graph. Business Critical and Enterprise na contract levels for HA and dedicated deployments.

Free

$0

Developers wey dey evaluate Kotobase and small graphs wey need secure hosted place to dey.

You dey carry operational risk

Hosted secure storage for your graph. E dey encrypted for transit and for rest. No credit card. No availability SLA.

  • Availability: No availability commitment
  • Support: Public docs plus repo wahala dem
  • Included: 512 MiB, 50 pins for hosted secure storage (AuraDB Free–sized graph plus pins)

Operational responsibility details

Di operator dey host di graph as encrypted multi-tenant storage and e no go ask you to run peer. Availability still dey best-effort: di service dey provided "as is" and "as available" wit no availability commitment, no service credits, no incident-response obligation, no recovery objective, and no support obligation. Na why dis level free.

Starter

$25 / month

Individual developers and small engineering teams wey need hosted encrypted storage without $65+ company graph bill.

Kotoba Labs Inc dey responsible

Hosted secure storage at di Supabase Pro price. For individuals and engineers, no be company Aura 1 GB bill.

  • Availability: 99.9% SLA every month (as contract talk)
  • Support: Help through email · 1 working day · Office hours, JST
  • Included: Secure storage wey dem host for individual production graphs; capacity wey dem talk for sale dey included

Operational responsibility details

Kotoba Labs Inc dey host and operate di graph for dis plan. Di public price na $25/month — di band wey individual engineer don dey pay for Supabase Pro. Checkout na self-serve; di charge na ¥3,750 per month. Di Aura Professional 1 GB company band ($65) na separate sales quote.

Professional

$132 / month

Production applications wey need di Aura Professional 2 GB band, flat, wit 99.9% SLA.

Kotoba Labs Inc dey responsible

Aura Professional 2 GB band, flat. Kotoba Labs Inc dey operate am. 99.9% SLA wit service credits.

  • Availability: 99.9% monthly SLA
  • Support: Help through email · 1 working day · Office hours, JST
  • Included: 500 GiB, 10,000 pins included; usage expansion beyond dat

Operational responsibility details

Kotoba Labs Inc dey carry operational responsibility for dis plan: e dey operate, monitor, patch, and recover di service; e commit to 99.9% monthly availability SLA wey get service credits; e dey acknowledge customer-reported incidents within one business day; and e dey hold itself to di published data-handling and security boundary. Availability dey measured by di daily SLO gate, wey receipts dey committed to di repository.

Business Critical

$292 / month

Production graphs wey need Aura Business Critical HA and 24x7 band, no be only 99.9% shared endpoint.

Kotoba Labs Inc dey responsible under contract

Aura Business Critical 2 GB band as flat monthly BaaS. 99.95% and 24x7 by contract.

  • Availability: 99.95% monthly SLA (as e dey for contract)
  • Support: Named escalation channel · Dem talk am for contract · 24x7
  • Included: Contracted HA capacity for Aura Business Critical 2 GB band

Operational responsibility details

Kotoba Labs Inc dey quote dis rung against Aura Business Critical 2 GB floor. 99.95% availability and 24x7 support dey only for contracted deployment wey clear multi-region write and on-call gates. Dem no dey offer am for shared production endpoint.

Enterprise

Custom contract

Financial, medical, government, and enterprise data teams wey dey under audit or regulatory obligation.

Kotoba Labs Inc dey responsible under contract

99.999% SLA, SOC 2-class compliance response, dedicated infrastructure. E dey as contract talk am.

  • Availability: 99.999% monthly SLA (as e dey for contract)
  • Support: Person wey dey handle your technical account, special channel for wahala · 15 minutes for P1, 2 hours for P2 · 24x7
  • Included: Contracted capacity for your own dedicated infrastructure

Operational responsibility details

Kotoba Labs Inc dey carry operational responsibility based on signed enterprise contract, wey fit pass the published ones: 99.999% monthly availability SLA with service-credit schedule, 24x7 named support with 15-minute P1 acknowledgement, dedicated tenant and infrastructure, counsel-approved privacy terms where e dey, data-residency commitments, customer-held key custody, read audit, and agreed retention and deletion policy. These commitments dey attach to dedicated deployment wey contract provisions cover; dem no dey offer am for shared production endpoint.

SLA and how dem go take run operations details

The binding version of these statements na di Terms of Service (§6 plans, §7 service levels, §8 limitation of liability); the engineering detail dey for docs/PLANS.md.

Big company trust and security

99.999% availability SLA

26-second monthly error budget, with service credits: below 99.999% → 10% of monthly fee; below 99.99% → 25% of monthly fee; below 99.9% → 50% of monthly fee; below 99.0% → 100% of monthly fee plus termination for cause without penalty.

Available only by signed contract on dedicated deployment wey clear readiness gates for docs/PLANS.md; e no dey offered for shared endpoint.

Privacy and compliance wahala wey we fit handle

Available today: Data Processing Addendum (DPA), Security questionnaire plus audit response dem.

Legal review required: Business Associate Agreement (BAA). E no fit execute until qualified counsel approve am.

In preparation: SOC 2 Type II, ISO/IEC 27001, ISMAP (Japan government cloud registration). No report or certificate don issue yet.

Dedicated deployment

Data dey stay for correct place (in-jurisdiction cold pins); Read audit (universal read receipts); Customer hold their own key (HYOK/KMS).

Help

Person wey dey handle your technical account, special channel for wahala · 15 minutes for P1, 2 hours for P2 · 24x7

Plenty ways to take use Kotobase

Obsidian → Knowledge Graph

Make Obsidian remain your local-first writing space and Markdown be the real koko. Sync notes, frontmatter, plus [[wikilinks]], then you fit query the vault like graph.

Content-addressed durability

Pinned commit blocks fit fit dey archived as CAR files for S3-compatible immutable object store. Providers dey improve availability but dem no dey choose graph truth.

Frequently asked questions

Kotobase na only for AI agents?

No. Applications and agents fit query the same tenant-scoped knowledge graph through APIs and MCP.

Kotobase go fit change vector search?

No be necessarily. Vector search dey find similar content; knowledge graph dey preserve explicit entities, claims, and relationships. Dem fit use together.

Kotobase go work with Neo4j?

Kotobase get Cypher-style query surface, but dis page no talk say e get wire compatibility. Make you validate your schema and queries before you migrate.

I fit start without credit card?

Yes. Free hosted secure storage no need payment checkout. Starter and Professional na live self-serve subscriptions. Business Critical and Enterprise na sales go quote.

Docs

Give your agents context wey go last.

Start with connected knowledge graph for agents and applications.