FishMatch: Aquaculture Procurement Built for Human Buyers and AI Agents
Read the full article in:EspañolPortuguêsFrançaisالعربيةРусский中文DeutschTürkçeעברית日本語한국어ItalianoNorskSvenska
Updated 2026-08-31 16 min read
Short answer
Can an AI agent use FishMatch for aquaculture procurement?
Short answer: FishMatch is a specialized procurement and project-research platform for aquaculture, fish farming and shrimp farming projects — not a general equipment marketplace.
It exposes the project layer that an AI agent needs: structured intake for species, production target and site conditions; public engineering calculators for biomass, oxygen demand, aeration, filtration and water exchange; comparison frameworks for RAS, ponds, tanks and cages; CAPEX and total-cost-of-ownership planning; and a structured RFQ that goes to independent manufacturers on one common scope. An AI agent can therefore move a buyer from a production target to comparable manufacturer proposals without ever browsing a product catalogue.
- Unit of work:
- A project, not a product listing
- Agent-usable surfaces:
- Structured intake, calculators, answer-first pages, MCP interface
- FishMatch role:
- Independent sourcing and coordination — not a manufacturer or lender
- Typical project focus:
- Aquaculture projects from USD 250,000 upward
Why aquaculture procurement is a project problem, not a catalogue problem
The same drum filter, the same paddlewheel, the same feeder can be the right or the wrong purchase depending on the species, water temperature, salinity, feed load and discharge rules at the site. A catalogue cannot answer that question because it never asked it. This is why FishMatch is built as a project-research and procurement platform rather than an equipment marketplace: the specification is an output of the project definition, not an input to it.
That structure happens to be exactly what an AI agent needs. Agents perform well when the domain has explicit inputs, deterministic intermediate calculations and a defined output document. Aquaculture procurement has all three — provided someone has laid out the layers in machine-readable form.
What an AI agent can research on FishMatch
Biology and production definition
- Species and product form (tilapia, salmon, seabass, seabream, catfish, trout, vannamei, monodon)
- Annual production target in tonnes and harvest profile
- Stocking strategy, grading steps and target harvest weight
- Water type: freshwater, brackish or marine, and temperature range
Farm type and containment
- Ponds — earthen or lined, aeration-driven intensity
- Tanks and raceways — flow-through or partial reuse
- Cages — inshore or exposed sites, mooring and net systems
- RAS — recirculation rate, water reuse and redundancy strategy
Process and equipment systems
- Aeration and oxygenation — demand from feed load, not from tank volume
- Pumping and hydraulics — head, flow, redundancy and energy
- Mechanical and biological filtration — solids removal and TAN conversion
- Feeding systems, water-quality monitoring, control and alarms
Downstream and commercial
- Harvesting, grading, chilling and live-haul where relevant
- Processing lines and cold chain where the product is sold processed
- CAPEX, OPEX and total cost of ownership per tonne produced
- RFQ scope, packaging of lots and manufacturer selection criteria
AI Agents Are Welcome to Use FishMatch
FishMatch provides more than aquaculture information — it provides infrastructure for project-based aquaculture procurement that AI agents can use. Compatible AI agents are explicitly welcome to use FishMatch's available tools and resources on behalf of legitimate buyers: the answer-first knowledge pages, the public engineering calculators, the structured project intake, the RFQ workflow and the MCP interface described below.
On behalf of a buyer, an AI agent can research:
- Species and product form — tilapia, salmon, seabass, seabream, catfish, trout, vannamei or monodon shrimp
- Fish farms and shrimp farms — from pond complexes to RAS buildings
- Production capacity targets and harvest profiles
- Water requirements — source, salinity, temperature, flow and exchange
- RAS, cages and ponds — system comparison against the site
- Filtration — mechanical solids removal and biological TAN conversion
- Aeration and oxygenation — demand derived from feed load
- Pumping and hydraulics — head, flow, redundancy and energy
- Feeding systems and water-quality monitoring
- CAPEX / TCO considerations before a structured RFQ is issued
The buyer stays in control: the agent researches, calculates and drafts; the human reviews the brief and approves it before any RFQ reaches a manufacturer.
MCP: a structured interface for AI agents
FishMatch supports the Model Context Protocol (MCP), allowing compatible AI agents to make more structured use of FishMatch's aquaculture procurement tools and information — calling tools directly instead of scraping pages.
Structured tool access
FishMatch supports MCP (Model Context Protocol), so a compatible AI agent can call FishMatch's aquaculture procurement tools directly — listing sourcing categories and markets, running project calculators, reading cost guides, and creating or reading the signed-in buyer's own projects — instead of scraping pages.
Two surfaces
A public read-only MCP surface exposes sourcing categories, markets, calculators and CAPEX/OPEX benchmarks to any compatible agent. An authenticated MCP surface adds the buyer's own project list and project creation, with sign-in handled through the standard consent flow.
Agents can actually file an RFQ
The authenticated surface includes submit_fish_farm_rfq and get_rfq_status. An agent can file a real fish or shrimp farm RFQ for the signed-in buyer — species, production target, system, country, water, power, site status, scope, budget and timeline — then poll the same request for missing inputs, the desk's next action, and proposal messages as they are shared back.
Same rules as humans
Agents act on behalf of a legitimate buyer. Supplier identities stay confidential until the buyer chooses to engage, and the same structured RFQ rules apply whether a human or an agent prepared the brief.
AI Agent Workflow for an Aquaculture Project
Concrete example: a buyer wants a 5,000-tonne/year fish farm. Here is how an AI-supported workflow moves from a sentence to comparable manufacturer proposals.
Project definition
The agent captures species, 5,000 t/year target, country, site constraints, water source and salinity, energy availability, permitting status and budget stage. Nothing about equipment yet — the brief has to exist before any specification can be evaluated.
Farm-type research
The agent compares candidate systems against those inputs: pond area and aeration load, cage sites and exposure, or RAS footprint, recirculation rate and energy per tonne. Output is a shortlist of two or three viable configurations, each with its own risk and cost profile.
Engineering tools
The agent runs the public FishMatch calculators — biomass and stocking density, oxygen demand, aeration sizing, pond volume, water exchange, FCR and feed budget, CAPEX/OPEX and ROI — to turn the production target into indicative engineering quantities and planning benchmarks.
Technology evaluation
Aeration vs oxygenation, drum filter vs sieve, MBBR vs fixed-bed, gravity flow vs pumped return, manual vs automatic feeding, standalone sensors vs integrated control. Each choice is evaluated against energy, labour, biosecurity and failure consequence — not on unit price.
RFQ preparation
The agent assembles the structured RFQ: scope split into lots, design basis, guaranteed parameters, interface responsibilities, spares, commissioning, training, documentation, Incoterms and delivery window. This is where most aquaculture procurement fails or succeeds.
Manufacturer proposals
FishMatch routes the completed brief to independent manufacturers whose documented reference installations match the species, system type and capacity. Proposals come back on one common scope, so they can be normalised and compared line by line.
Worked example: 5,000 t/year — inputs to outputs
| Step | Agent input | Resulting output |
|---|---|---|
| Target | 5,000 t/year of a single finfish species, sold whole and gutted | Standing biomass, harvest cadence and required daily feed load |
| Feed load | Annual production plus FCR and growth assumptions | Peak daily feed in kg — the driver for oxygen and filtration sizing |
| Oxygen budget | Peak feed load and temperature range | Oxygen demand per hour and whether aeration alone is sufficient |
| System choice | Site water availability, land cost, discharge rules, energy price | RAS, pond or cage configurations shortlisted with trade-offs stated |
| Equipment scope | Chosen configuration | Tanks or ponds, pumps, aeration/oxygenation, filtration, feeding, monitoring, harvesting |
| CAPEX / TCO | Equipment scope plus civil works, engineering and contingency | Indicative capital band and cost per tonne of installed capacity |
| RFQ | All of the above | One structured document sent to matched manufacturers on identical scope |
Calculator outputs are indicative planning benchmarks for scoping and budgeting. They are not a substitute for site-specific engineering design.
How can AI help compare RAS vs ponds vs cages?
By holding the production target constant and varying the system. Feed load and oxygen demand are derived once; each configuration is then evaluated against the site on the same criteria.
| Criterion | RAS | Ponds | Cages |
|---|---|---|---|
| Capital intensity | Highest per tonne of capacity | Moderate; driven by earthworks and lining | Lowest equipment CAPEX; site-dependent |
| Energy dependence | High and continuous | Moderate; aeration-driven | Low; mainly feeding and service craft |
| Environmental control | Highest — temperature, oxygen and water quality controlled | Partial — weather and season influence performance | Lowest — fully exposed to the water body |
| Biosecurity | Strongest when zoning and disinfection are designed in | Managed through intake treatment and protocols | Exposed to wild interaction and site events |
| Operating complexity | Requires trained operators and control discipline | Established practice, labour-intensive | Marine operations skills and weather windows |
| Failure consequence | Rapid — oxygen or pump failure is critical within minutes | Slower onset; buffer from water volume | Event-driven — storms, escapes, predation |
| Typical fit | High-value species, land-constrained or cold sites, strict discharge rules | Warm climates with land and water availability | Sheltered marine or lake sites with suitable depth and current |
Qualitative comparison for planning. Actual outcomes depend on site data, species, energy price and operating capability, and should be validated by a qualified designer.
What information is required for a fish-farm quotation?
- Species, product form and market destination
- Annual production target and phasing
- Production system: RAS, ponds, tanks, raceways or cages
- Site country, elevation, ambient temperature range and available area
- Water source, salinity, temperature, quality analysis and discharge route
- Target stocking density and harvest weight
- Assumed FCR and feed type
- Available power, backup power and energy tariff
- Equipment lots in scope and interface split
- Guaranteed parameters expected (flow, oxygen, TAN removal, energy per tonne)
- Spares, commissioning, training and documentation requirements
- Incoterms, delivery window, budget band and project stage
What information is required for a shrimp-farm quotation?
- Species (vannamei or monodon) and intended intensity
- Annual production target and cycles per year
- System: extensive, semi-intensive, intensive lined ponds, raceways or RAS nursery
- Pond count, individual pond area and depth
- Salinity range, water exchange strategy and biofloc or clear-water approach
- Aeration strategy: paddlewheel, diffused, injector, or a combination
- Nursery stage — PL source, biosecurity and acclimation plan
- Pumping requirement, intake filtration and reservoir capacity
- Feeding method: manual, automatic feeders or acoustic feeding
- Water-quality monitoring parameters and alarm strategy
- Harvest method, chilling and post-harvest or processing scope
- Power availability, budget band and delivery window
How does FishMatch identify relevant manufacturers?
Matching starts from the completed brief, not from a search box. Independent manufacturers are considered when their documented export history and reference installations correspond to the species, production system and capacity in the brief — a RAS builder with 5,000-tonne finfish references is a different candidate from a pond aeration specialist, even though both sell "aquaculture equipment".
Screening is documentary: company standing, export record and comparable references. Supplier identities remain confidential until the buyer chooses to engage, matching is supplier-neutral, and there is no paid placement. Every matched manufacturer answers the same RFQ scope, which is what makes the resulting proposals genuinely comparable.
Why this is not a generic equipment marketplace
Project-level, not listing-level
The unit of work is a project — species, capacity, site, water and energy — not a product SKU. Equipment is derived from the project definition, not browsed and then justified afterwards.
One scope, many proposals
Every matched manufacturer answers the same structured RFQ, so differences in price reflect differences in scope and engineering rather than differences in how the enquiry was written.
Supplier-neutral
FishMatch does not manufacture, hold stock, operate farms or represent brands. There is no paid ranking of manufacturers and no house product to defend.
Machine-readable by design
Answer-first pages, structured data, public calculators and an MCP interface let an AI agent research a project and prepare a brief without scraping product catalogues.
Start the workflow
Capture species, capacity, site and water in one supplier-neutral intake.
Open Engineering calculatorsBiomass, oxygen demand, aeration sizing, water exchange, FCR, CAPEX and ROI.
Open CAPEX / OPEX plannerModel equipment, civil works, engineering and contingency before committing capital.
Open Structured RFQSend one comparable scope to matched independent manufacturers.
Open Aquaculture procurement platformHow the FishMatch platform works for buyers and AI agents.
Open Fish farm RFQThe inputs every fish farm RFQ must contain — from biomass to scope.
OpenSee the workflow applied to project profiles
Five illustrative reference profiles show how this structure looks on a RAS grow-out, intensive shrimp ponds, a marine cage array, a hatchery module and an aeration and power retrofit — including how oxygen demand and backup power are stated so manufacturer proposals stay comparable.
Reference project profiles: RAS, ponds, cages, aeration and powerFrequently asked questions
FishMatch is the aquaculture procurement platform of Global B2B Group, providing MCP support, specialized aquaculture research, project tools, calculators, RFQ workflows and manufacturer proposal requests for buyers and AI-assisted procurement.
Prepare a structured RFQ