/usr/bin/graphsub — 132×43 — us-east-1 --:--:-- UTC · xterm-256color · UTF-8
BOOT[BOOTING]

graphsub@cloud:~$ graphsub --describe

Built for agents.

Sub-millisecond graph operations, stigmergic coordination, hierarchical memory, and replayable proof for systems where agents and operators need one operational truth graph.

67µsP99 lookup
25Mops/sec
1 graphmemory + proof
130× cost-effective 166× faster Native agent primitives Fundable capsules Live product console
initialize_free_tier

50K GSU included · No credit card required · press / for commands

LIVE://TELEMETRY● sim
regionus-east-1
agents.live1,214,069
graph.nodes824,326
graph.edges3,785,778
lookup.p9967µs
throughput24.7M ops/s
OPS/SEC24.7M OPS/S
┤ FIELD://SUBSTRATE · cursor deposits · traces become structure ├
EVENTS://STREAM● sim
SSE · /api/v1/events/stream0 ev/shover to pause
SHARDS://256LSM · RAFT · SWIM
OP MIX
MEM TIERSw:0 · e:0 · s:0
SESSION GSU0 · $0.00
67µs
Node Lookup
P99 LATENCY
25M
Operations
PER SECOND
6
Capsules
PILOT WEDGES
131
API Endpoints
REST + GRAPHQL
3-tier
Memory
HIERARCHICAL

graphsub@cloud:~/graphsub$ ls -l ./capabilities

01

Native agent primitives

8 MODULES · SELECT TO INSPECT
MODULES
./CAPABILITIES/MEMORYLOADED

graphsub@cloud:~/cap/memory$ cat manifest.json

Hierarchical Memory

Working → Episodic → Semantic. Automatic consolidation with configurable decay.

graphsub@cloud:~/graphsub$ graphsub examples --capsules --syntax

01B

Code-first capsules

PYTHON · CYPHER · REST · REPLAY
PYTHON://AGENT_MEMORYCOPYABLE PATTERN
memory_replay.pywrite · recall · replay
from graphsub import Client

client = Client(base_url="https://graphsub.com")

memory = client.memory.remember(
    tier="working",
    content="limit breach on EUR rates book",
    source="risk-run-2026-07-03",
    importance=0.92,
)

trace = client.capsules.execute(
    capsule="trading-control-graph",
    workload="risk-recompute",
    evidence="source-linked",
)

print(trace.replay_id, trace.affected_outputs)

The pitch is concrete: memory write, capsule execution, affected-output replay, and evidence links in one path.

CYPHER://DEPENDENCY_GRAPHAGENT STATE
affected_outputs.cypherblast radius query
// Which reports need recompute after a curve update?
MATCH (s:Source {id: 'SOFR:2026-07-03'})
  -[:FEEDS]->(m:Model)
  -[:PRODUCES]->(o:Output)
WHERE o.status <> 'replayed'
RETURN
  o.id AS output,
  m.version AS model,
  shortestPath((s)-[*..4]->(o)) AS proof_path
ORDER BY o.priority DESC

This is the product wedge: explain impact, recompute only what changed, and keep the proof path inspectable.

REST://CAPSULE_PLANLIVE MODAL ROUTE
capsule_plan.sh/api/plan · /api/execute · /api/examples
curl 'https://graphsub.com/api/plan?capsule=trading-control-graph&target=enterprise-team&workload=risk-recompute' \
  | jq '.execution_preview | {capsule, buyer, workflow, artifacts}'

curl -X POST 'https://graphsub.com/api/execute?capsule=regulated-evidence-graph&workload=audit-evidence' \
  | jq '.execution.replay'

# crawlable examples for LLMs, docs, and sales engineering
curl 'https://graphsub.com/api/examples'

The landing page now exposes implementation-shaped proof, not only market copy.

graphsub@cloud:~/graphsub$ graphsub repl

02

The API surface

131 ENDPOINTS · VERIFIED TABS RUN LIVE CAPTURES
REPL 

        

          
ENDPOINT

          

          
          

          

        

        
// output — press RUN to replay the selected request.

      

    

  







  

    
graphsub@cloud:~/graphsub$ graphsub console --launch

    
02B
Live product console
CAPSULE PLANNER · BENCHMARK API · MARKET MAP

    

      

        CONSOLE://LIVECONNECTING
        
Launch a graph-native workflow

        
Loading live product data.

        

          
--capsules

          
--competitors

          
--bench lanes

        

        

      

      

        BUILDER://CAPSULE_TO_WORKFLOWPOST /api/execute
        
Build a capsule workflow

        
Pick the workflow, buyer, and output mode. GraphSub returns a launch plan with schema, replay, benchmark, and ROI outputs.

        

          
          
          
          
        

        

          
          
        

        
awaiting /api/plan ...

      

    

    

      
/api/plan
workflow plan
Capsule launch steps with selected buyer and workload.

      
/api/evidence
source links
Competitor and benchmark context with direct source routes.

      
/api/execute
launch plan
Manifest, schema, replay, benchmark slice, and ROI worksheet.

    

  







  

    
graphsub@cloud:~/graphsub$ graphsub capsules --fundable --real-world

    
03
Fundable capsules
BUYER → PROOF → PLATFORM

    

      
CAPSULES

      

        CAPSULE://TRADING_CONTROL_GRAPHPILOT-READY
        
graphsub@cloud:~/capsules$ inspect trading-control-graph

        
Trading Control Graph

        

        

          
BUYER

          
WEDGE

          
PILOT

          
COMMERCIAL

        

        

          
PROOF ARTIFACT

          
DATA PLANE

          
MOAT

          
EXPANSION PATH

        

        

          
FUNDABILITY SIGNAL

          

        

      

    

    

      

        
DILIGENCE BINDER

        
Each capsule ships with architecture map, data contracts, trace log, eval harness, and ROI worksheet.

        inspectable, not narrative-only
      

      

        
PILOT PROTOCOL

        
One workflow, one buyer, one success metric, one replayable output. Start narrow, then expand into the control graph.

        30-day shadow run
      

      

        
DEFENSIBILITY

        
Subgraph deltas, blessed outputs, scenario scopes, dependency invalidation, and trace hooks compound by deployment.

        customer semantics become the map
      

      

        
GO-TO-MARKET

        
Land where graph errors are expensive, then expand into memory, evals, audit, and autonomous operations.

        operator pain before platform pitch
      

    

    

      

        90-DAY MOTION
        
Capsule-to-company plan

        
    
          
  • days 0-14Pick one capsule, import a narrow real workflow, define blessed outputs and failure modes.
    • 
          
  • days 15-30Run a shadow pilot with replay, invalidation, and proof artifacts against the incumbent process.
    • 
          
  • days 31-60Turn the pilot into a repeatable capsule pack: schema, demo data, eval suite, security memo, buyer deck.
    • 
          
  • days 61-90Close design partners, price the control plane, and prove expansion from one graph wedge to platform usage.
    • 
        

      

      

        INVESTOR THESIS
        
Why this can be venture-scale

        

          
Agent era: autonomous systems need persistent, queryable, replayable memory and tool traces.

          
Enterprise wedge: regulated and complex operators already pay for control, audit, and recomputation correctness.

          
Platform pull: each capsule starts narrow, then pulls in graph storage, evals, security, and orchestration.

          
Proof posture: every capsule produces replayable outputs, source links, and a clear success metric.

        

      

    

  







  

    
graphsub@cloud:~/graphsub$ graphsub market --compare puppygraph,neo4j,memgraph,tigergraph,neptune

    
04
Competitive market map
OFFICIAL SOURCES · JUL 2026 REFRESH

    

      
VENDORS

      

        COMPARE://PUPPYGRAPHSOURCE-LINKED
        
PuppyGraph

        

        

          
BEST AT

          
STACK POSTURE

          
QUERY SURFACE

        

        

          
WHY THEY WIN

          
GRAPHSUB MOVE

        

        

          
AXIS SCORECARD

          

        

        

      

    

    

      MATRIX://DO_NOT_FIGHT_COMMODITY_GRAPH_WARS
      

    

    

    

      PUBLIC_STATS://COMPETITOR_EVIDENCEPUBLIC SOURCES · SOURCE-LINKED
      

      
Competitor context is linked to public sources so teams can compare architecture, query model, and deployment posture quickly.

    

  







  

    
graphsub@cloud:~/graphsub$ graphsub bench --compare all --fast · graphsub market --stats all

    
05
Expanded benchmark map
MEASURED LANES + PUBLIC CLAIM LANES

    

      BENCH[IDLE]
      

      
all competitors shown · GraphSub performance lane · peer and market context lanes

    

    

    

      BENCH_ALL://FAST_PUBLIC_NORMALIZERMEASURED + CLAIMED LANES
      

      
Use this map to understand where GraphSub wins: low-latency graph operations, agent memory, replay, and capsule workflows.

    

  







  

    
graphsub@cloud:~/graphsub$ graphsub openontology --tree · objects, links, actions, functions, scenarios, and permissions

    
06
OpenOntology on boot
65,536 CONCEPTS · 512-TYPE KERNEL · REALTIME

    

      
ONTOLOGY://TREE
        

      

      
ALGORITHMS://PAGERANK[LIVE CAPTURE]
        
$ pagerank --top 8 · POST /api/v1/shards/0/algorithms/pagerank

        

        
OpenOntology binds object types, link types, shared properties, actions, functions, interfaces, object sets, scenarios, provenance, permissions, and agent runtime state. 8 algorithms live: pagerank · betweenness · closeness · katz · shortest_path · all_shortest_paths · community_detection · connected_components.

      

    

  







  

    
graphsub@cloud:~/graphsub$ graphsub billing --explain

    
07
One meter. No surprises.
GSU · $0.10 PER 1,000

    

      GSU://CONVERSIONS
      
ALL USAGE CONVERTS TO GSU (GraphSub Units) — ONE METRIC ACROSS CYPHER · REST · GRAPHQL$0.10 / 1,000 GSU

      

        
1 Node
1 GSU

        
1K Ops
10 GSU

        
1K Traversals
2 GSU

        
1 Vector Query
5 GSU

        
1K Analytics
25 GSU

        
1 LLM Extract
50 GSU

      

    

    

      

        
        
FREE

        
$0 forever

        
    
          
  • 100K nodes
    • 
          
  • 50K GSU / month
    • 
          
  • All features
    • 
          
  • Community support
    • 
          
  • No credit card required
    • 
        

        initialize_free_tier
      

      

        ▸ MOST POPULAR
        
PAY-AS-YOU-GO

        
$0.10 / 1K GSU

        
    
          
  • Unlimited nodes + GSU
    • 
          
  • 100 req/s
    • 
          
  • 3 projects
    • 
          
  • Email support
    • 
        

        start metering
      

      

        ▸ BEST VALUE
        
COMMITTED USE

        
−25% volume

        
    
          
  • Discounts at $10K / $25K / $50K / $100K per month
    • 
          
  • Dedicated support
    • 
          
  • Custom limits
    • 
        

        talk to us
      

    

    
Teams need SSO, audit, SLAs? Platform seats: Pro $99/seat · Enterprise $499/seat — docs.

  







  

    
Ready for the agent era?

    
Start free. Scale to millions. No credit card required.

    
graphsub@cloud:~$ initialize_free_tier --now

    initialize_free_tier