Milestone 1 Verified · 100k+ Sovereign Literature Grounded

Every species.In context.

Automated extraction and grounding of ecological interactions with 100% literature sentence traceability. Over 100,000 sovereign interaction records verified against PubMed Central, resolving scientific names across NCBI Taxonomy and GBIF Backbone in sub-millisecond latencies.

Explore 100k Records Try HF Space Demo ⚡ Human Validation Bench

100,128 TIER 1 LIT-GROUNDED RECORDS·23,997 LINNAEAN TAXA·B-TREE <50µs·PARQUET SHA-256 SEALED

100,128+

Tier 1 Sovereign Interactions (100% Sentence Grounded)

145,851

Total Knowledge Graph Edges (Tier 1 + Tier 2)

<50µs

B-Tree SQLite Entity Resolution

8,341

Peer-Reviewed Open Access PMC Articles

*Local backend execution. Live request timings include network roundtrip and Cloudflare proxying.

01 / Interactive explorer

Taxon resolution, grounding & triage.

Test scientific name normalization, document context grounding, and abstract interaction screening with live API execution.

Live API sandbox

A name is just the beginning.

Resolve a taxon to discover its accepted name, checklist identifiers, and synonym relationships.

View raw JSON payload

Small abbreviations. Full context.

See how B. terrestris connects to the full species name introduced in the abstract.

View raw JSON payload

Article title TITLE

Abstract text TRIAGE TARGET

Pre-configured presets

Sub-millisecond screening (<150 µs)

Ultra-fast screening before deep extraction.
EcoTriage rejects non-ecological literature (pure clinical, medical, cellular) in microseconds so downstream LLM and graph extractors never waste compute.

Screen literature in microseconds.

Paste an abstract to instantly test ecological interaction probability, detected cues, and candidate taxa.

View raw JSON payload

Local GBIF 7.7M Backbone NCBI Taxonomy MDD v2.5 EcoTriage <150 µs Explore API OpenAPI Spec ↗

Local 7.7M Backbone

Instant SQLite indexed taxonomic resolution across plants, fungi, animals, and microbes with zero remote API failure rate.

Document Grounding

Connect shorthand notations in Results (e.g. B. terrestris) to full taxonomic inventories defined in the Abstract.

Microsecond Triage

Pre-filter thousands of scientific papers per minute, routing only true ecological interaction literature to extraction pipelines.

02 / Ecosystem in action

EcoInteraction Knowledge Explorer.

Live queryable knowledge base adhering to IS-1.0 and OBO Relations Ontology. Harvested continuously from open-access scientific literature and curated datasets.

v1.0 Sovereign Parquet (3.6 MB) SHA-256: 760446bc...

Live Knowledge Graph Explorer

Over 100,000 sovereign interaction edges grounded against GBIF Backbone 7.7M + NCBI Taxonomy with 100% verbatim text verification.

Total Edges: 145,851

Tier 1 Sovereign: 100,128

Linnaean Taxa: 23,997

Source Taxon (Subject)	Interaction (OBO RO)	Target Taxon (Object)	Valence	Literature Grounding	Details

Showing 1–25 of 100,128 sovereign edges

Page 1 of 476

03 / Under the canopy

Three checklists. A clearer picture.

Taxonomy evolves. EcoTagger retains the evidence behind every name.

Names connected, differences preserved.

EcoTagger brings independent taxonomic views into one response. Trace a name to its identifiers, inspect synonymy, and keep checklist disagreements explicit.

Mention → Resolver → Local GBIF Backbone + NCBI + MDD → Grounded Consensus

A shared identifier for biological names

NCBI Taxonomy connects scientific names to TaxIDs used throughout sequence databases. An NCBI match can remain useful even when a name is absent from another checklist.

ncbi_taxid: "9696"

Local 7.7 Million Backbone (SQLite)

GBIF returns a usage key, accepted name, rank, and match status. Powered by a local SQLite cache containing 7,725,567 taxa for microsecond response times with remote API fallback.

gbif_usage_key: 2435104 · status: "SYNONYM" · source: "local_sqlite_backbone"

A specialist view of mammalian diversity

MDD v2.5 adds a mammal-specific checklist relation. An ABSENT result means there is no match in this checklist; plants, insects, and viruses are outside its scope.

relation: "SYNONYM_OF" · resolved_name: "Puma concolor"

ACTIVE EXPANSION ROADMAP Extensible Multi-Taxonomic Core

Broadening beyond NCBI / GBIF / MDD

EcoTagger's resolver architecture is fully decoupled from any single taxonomic authority. To span the full global tree of life across marine, terrestrial, and microbial realms, we are actively onboarding authority checklists into our consensus resolver:

Catalogue of Life (CoL) In Progress

Global comprehensive consensus checklist indexing over 2.1 million living species.

Open Tree of Life (OTT) In Progress

Phylogenetic synthetic tree integrating 2.4M taxa for evolutionary context.

WoRMS Scheduled

World Register of Marine Species, authoritative standard for oceanic biodiversity.

ITIS Scheduled

Integrated Taxonomic Information System (TSN identifiers for North American & global taxa).

MycoBank & Index Fungorum Scheduled

Authoritative nomenclatural repository for fungal and lichenized species.

Built for a different research question.

EcoTagger / NCBI PubTator 3.0

Scope and design comparison of EcoTagger and NCBI PubTator 3.0
Capability	EcoTagger	NCBI PubTator 3.0
Research focus	Ecology, interactions & biodiversity literature	Biomedical literature & semantic search
Taxonomic views	NCBI + Local GBIF 7.7M + MDD v2.5	NCBI Taxonomy species normalization
Document context	Title & Abstract ground Results abbreviations	Entity annotations in PubMed & PMC full text
Checklist disagreements	Explicit relations & conflict labels	No equivalent multi-checklist response
Literature Triage	Built-in EcoTriage (<150 µs pre-filter)	Medical indexing without interaction gating
Relation vocabulary	22 ecological types in IS-1.0	12 biomedical relation types

A comparison of scope, not an accuracy benchmark.

04 / Developer integration

One request.
A world of context.

Integrate taxon resolution, entity grounding, and microsecond interaction triage into your own Python, R, or Bash pipelines.

Swagger UI (/docs) ReDoc ↗

# 1. Resolve a synonym across checklists
curl -X POST https://ecotagger.ecoseek.org/v1/resolve \
  -H 'Content-Type: application/json' \
  -d '{"name": "Felis concolor"}'

# 2. Ultra-fast interaction triage (<150 µs)
curl -X POST https://ecotagger.ecoseek.org/v1/triage \
  -H 'Content-Type: application/json' \
  -d '{"title": "Puma predation", "abstract": "Puma preys on deer."}'

import requests

# Resolve taxon
res = requests.post(
    "https://ecotagger.ecoseek.org/v1/resolve",
    json={"name": "Felis concolor"},
    timeout=30,
)
print(res.json()["scientific_name"])  # Puma concolor

# Screen abstract with EcoTriage (<150 µs)
triage = requests.post(
    "https://ecotagger.ecoseek.org/v1/triage",
    json={"title": "Predation dynamics", "abstract": "Puma concolor preys on deer."},
)
print(triage.json()["decision"])  # True

library(httr2)

# Resolve taxon
resp <- request("https://ecotagger.ecoseek.org/v1/resolve") |>
  req_body_json(list(name = "Felis concolor")) |>
  req_perform()

data <- resp_body_json(resp)
print(data$scientific_name)

# EcoTriage pre-filter
triage <- request("https://ecotagger.ecoseek.org/v1/triage") |>
  req_body_json(list(title = "Predation dynamics", abstract = "Puma preys on deer.")) |>
  req_perform() |>
  resp_body_json()
print(triage$decision)

Every species.In context.

Taxon resolution, grounding & triage.

A name is just the beginning.

Small abbreviations. Full context.

Screen literature in microseconds.

Local 7.7M Backbone

Document Grounding

Microsecond Triage

EcoInteraction Knowledge Explorer.

Live Knowledge Graph Explorer

Three checklists. A clearer picture.

Names connected, differences preserved.

A shared identifier for biological names

Local 7.7 Million Backbone (SQLite)

A specialist view of mammalian diversity

Built for a different research question.

One request.A world of context.

One request.
A world of context.