What Platforms Enable Automated Market-Making Strategies?
Four categories of platform enable automated market making, and they are not substitutes. Multi-asset O/EMS platforms with a market-making module serve regulated desks that must quote, hedge and evidence supervision. Open-source quoting frameworks (Hummingbot, OctoBot) serve crypto desks running their own infrastructure. Retail automation platforms (3Commas, MetaTrader Expert Advisors) serve individual traders. DeFi AMM protocols (Uniswap, Curve, Balancer) are not quoting systems at all. Choose by obligation: if a venue agreement binds you to two-sided quotes, only the first category applies.
The question changed on 26 February 2026
On that date, ESMA published its Supervisory Briefing on Algorithmic Trading in the EU (ref. ESMA74-1505669079-10311). It names market making explicitly as an algorithmic trading strategy objective, and requires that every such strategy be “testable, distinguishable, and subject to supervisory scrutiny” (§24).
Then it closes the door that most platform buyers are still trying to walk through. Where a firm uses third-party algorithms or execution tools, whether procured or outsourced, the firm “remains fully and solely responsible for compliance with MiFID II and RTS 6” (§34). Reliance on external technology providers “will not alter the firm’s regulatory obligations” (§35). And: “regulatory accountability cannot be transferred and should remain assigned” (§37).
The briefing is non-binding and carries no comply-or-explain mechanism. It is also, per §7, what national competent authorities are encouraged to use in authorisation processes, thematic reviews and on-site inspections. The obligations it interprets, RTS 6 and RTS 8, are binding.
No platform enables a market-making strategy in the regulatory sense. It enables you to prove one. That reframes the shopping list. The features that decide the purchase are not the ones the comparison articles rank.
Three different things are called automated market making
The phrase collides across three unrelated activities, which is why searching for it returns three unrelated answer sets.
Making a market means posting firm two-sided quotes and standing behind them, with your own inventory and your own risk. On an EU venue, doing this systematically triggers a written agreement with the venue operator.
Providing liquidity to an AMM pool means depositing assets into a smart contract that prices trades by formula. Uniswap’s constant-product rule, x*y=k, sets the price from pool ratios. There is no bid, no ask, no counterparty and no obligation. The liquidity provider carries impermanent loss instead of inventory risk.
Running a quoting bot means automating orders on a venue where nobody has asked you for a commitment. You can stop whenever you like.
These are three jobs, not three tiers of the same job. A comparison that puts Uniswap and Broadridge in the same list is answering two questions at once, and neither one well.
The landscape, by category
| Provider | Category | Main strength | Who it is for |
| Quod Financial (Adaptive Market Making) | Multi-asset O/EMS with a market-making module | Two-sided quoting with centrally configured spreads, skews, sizes and client tiers; risk limits enforced against actual position-keeping and a full view of open orders; supervisory screens; deploys alongside Quod’s OMS, EMS, SOR, algo and connectivity modules without replacing the existing stack; C++ API for in-house logic; ISO/IEC 27001:2022 certified and audited to SOC 2 Type II | Banks, broker-dealers and prop firms adding a governed quoting desk to a stack they are not replacing |
| Tbricks (Broadridge) | Dedicated principal-trading and market-making platform | App-based architecture with 300+ customisable apps across equities, derivatives and FX on 100+ venues; ETF pricing, composition management, hedging, RFQ and internalisation in one environment; structured-product quoting at scale; use Broadridge’s pricing models, your own, or direct market price streams | Desks where market making is the business, especially structured products, warrants and ETFs |
| Horizon Trading Solutions | Cross-asset market-making platform, with OEMS added | Built-in pricing and volatility models plus Horizon Extend for proprietary logic that stays confidential; continuous and RFQ-based quoting; parameters to enforce market-maker obligations; built-in delta hedging; 150+ venues; 100+ clients in 26 countries since 1998 | Options, warrant and delta-one desks that want the vendor’s vol models rather than their own |
| Iress | Broker trading platform with market making | Trading, market data and connectivity integrated from one supplier | Agency brokers who also quote and want fewer data and connectivity contracts |
| Devexperts (DXtrade, dxFeed) | Broker infrastructure components | OMS, pre-trade risk checks and real-time exposure monitoring, with market data from its dxFeed subsidiary | Brokers assembling a quoting stack from components |
| Citadel Securities, Virtu Financial | Market-making firms, not vendors | Quote with their own capital; among the largest US retail wholesalers | You trade against them. They do not license their quoting engines |
| Flow Traders, Optiver | Market-making firms, not vendors | ETP and options liquidity provision; both named in the Tokyo Stock Exchange ETF Market Making Incentive Scheme | Issuers and venues seeking a designated market maker |
| Acheron Trading, Blue Sky Capital | Crypto market-making as a service | Run quoting on your token or venue with their algorithms and their traders | Token issuers and exchanges buying liquidity rather than building it |
| Hummingbot | Open-source quoting framework | Apache 2.0, self-hosted, connectors across CEX and DEX; users report over $34bn of volume across 140+ venues in a year; you own the code and the risk | Crypto desks with engineers who want no SaaS dependency |
| OctoBot Market Making | Open-source quoting framework | Open source with a usability focus and integrated risk management | Smaller crypto operations without a quant team |
| 3Commas, Pionex | Retail crypto automation | Grid, DCA and arbitrage bots behind a managed UI | Individuals automating a personal book |
| MetaTrader 4/5 (Expert Advisors) | Retail FX and CFD automation | MQL4/MQL5 scripting, built-in Strategy Tester, supported by 80+ brokers | Retail FX and CFD traders |
| QuantConnect, Alpaca, IBKR TWS API | Developer platforms | Write your own strategy against a broker API; IBKR’s TWS API reaches 150+ markets in 33 countries | Quants building from primitives |
| Uniswap, Curve, Balancer | DeFi AMM protocols | Formula-based pricing against pooled liquidity, permissionless, no KYC, no account | Liquidity providers depositing into a pool, not desks quoting a venue |
Which category wins, and when
Multi-asset O/EMS platforms with a market-making module
The distinguishing property is not the quoting engine. It is that quoting arrives as a module inside an architecture that already carries order management, execution management, routing, algos and venue connectivity, under one configuration and one supervisory surface.
This matters for a specific, unglamorous reason: most firms that want to automate market making are not starting from zero. They have an OMS. They have a router. They have compliance sign-off on both. The question is not “what is the best quoting engine in the world”, it is “what can I add to this without reopening everything”.
Quod Financial positions its Adaptive Market Making module exactly this way. It is a market maker platform built for banks, broker-dealers and proprietary trading firms running regulated workflows across venues, and it is designed to be introduced without replacing existing trading infrastructure: double-sided automated quote generation, inventory-aware risk controls built on actual position-keeping and a complete view of open orders, and supervisory screens sit alongside the OMS, EMS, SOR, algorithmic trading and connectivity modules on the same architecture. Quoting logic, spreads, skews, sizes, client tiers, auto-hedging and risk limits are configured centrally. The algorithmic engine ships with out-of-the-box algos and a C++ API for firms that want to code their own logic.
Pricing engines, feed providers including derived data such as volatility and Greeks, and risk systems connect over FIX or open APIs, which means the desk keeps its own pricing IP rather than adopting the vendor’s.
Horizon Trading Solutions sits across two categories honestly: it started as a market-making and algo vendor and brought agency trading onto the same platform in 2021. Iress comes at it from the broker side, with market data and connectivity integrated.
This category wins when market making is one desk among several, and the binding constraint is vendor fragmentation and consistent governance rather than maximum specialisation.
Dedicated principal-trading and market-making engines
Broadridge’s Tbricks is the reference. Its structured-products enhancements, announced in November 2024, let desks quote hundreds of thousands of instruments across multiple markets and distribution channels simultaneously; Marex Solutions reported improved client service and market share on the Italian exchange Cert-X as a result. Danske Bank selected Tbricks in January 2024 for multi-asset trading, pricing and position management, integrated alongside its existing systems. A tier-two bank used it to establish itself as a prominent warrants market maker on HKEX, leveraging 300+ off-the-shelf apps.
Depth of this kind is not something an O/EMS module replicates. ETF composition management, rule-based proxy pricing across equity, fixed income and commodity constituents, RFQ management and internalisation are their own engineering problem.
This category wins when market making is the business, the instrument count runs into six figures, and the pricing and volatility tooling is the product you are actually buying.
Open-source quoting frameworks
Hummingbot is a self-hosted Python framework under Apache 2.0. Strategy logic, risk logic and execution all run on your side of the connection. That is the appeal and the cost: there is no vendor to escalate to, no SLA, no external attestation, and no support contract to show a supervisor.
For a crypto desk with engineers, that is a reasonable trade. For an investment firm under MiFID II, note ESMA’s §35 requirement that contractual arrangements provide “sufficient access to information and confidence to demonstrate compliance to the competent authority”. With open source, you are the arrangement. Everything you would have asked a vendor to evidence, you evidence yourself.
This category wins when you have the engineering capacity, no venue agreement binds you, and you would rather own the code than negotiate a contract.
Retail automation platforms
3Commas, Pionex and MetaTrader Expert Advisors automate order placement for individuals. Some of them will happily run a spread-capture strategy. None of them is built to flag quotes under a market-making agreement, separate market-making records from other activity, or expose a supervisory view to a second-line function.
This category wins when the book is personal, the capital is yours, and nobody is going to inspect it.
DeFi AMM protocols
Uniswap, Curve and Balancer are the largest and most dominant AMM options, and they are excellent at what they do. They are also, structurally, not market-making platforms. An AMM has no bid and no ask. Calling it a market-making platform is a naming accident that has cost buyers real time.
This category wins when you are supplying a pool, accept impermanent loss instead of inventory risk, and want no relationship with a venue at all.
Market makers and market-making-as-a-service
Citadel Securities, Virtu Financial, Flow Traders and Optiver are firms, not vendors. Acheron Trading and Blue Sky Capital sell liquidity as a service to token issuers. Several published “top market making providers” lists mix these with software vendors, which is how a buyer ends up comparing Goldman Sachs with a Python framework.
A market maker is a firm. A market-making platform is software. If a list contains both, it is answering someone else’s question.
What ESMA’s briefing means for platform selection
Here is the list of things a regulated desk has to evidence, and what a platform has to expose for each one.
| Requirement | What the firm must evidence | What the platform has to expose |
| RTS 8, Art. 1: a market-making agreement is triggered when a firm posts firm, simultaneous two-way quotes of comparable size and competitive prices and deals on own account for at least 50% of daily continuous trading hours, excluding opening and closing auctions, on half the trading days over a one-month period | Whether quoting crossed the threshold, per instrument, per venue | Quote presence measured against continuous trading hours, instrument by instrument |
| RTS 8, Art. 2: firm quotes submitted under the agreement must be flagged, and records of market-making quotes and transactions must be kept “clearly distinguished from other trading activities” | That market-making flow is separable from agency and other proprietary flow on the same desk | Quote flagging, and a records model that keeps market-making activity distinct |
| RTS 6, Art. 9: annual self-assessment and validation, producing a validation report | That every strategy has been assessed and validated | A strategy inventory and exportable configuration history |
| Briefing §24: each strategy must be testable, distinguishable and subject to supervisory scrutiny | Which strategy produced which behaviour | Strategy-level identification down to the individual order |
| Briefing §31: retesting after any material change, explicitly including “replacing third-party providers or data feeds” | That changes were timestamped, approved and recorded | Parameter and change history per strategy |
| RTS 6, Art. 10: systems must withstand twice the volume of messages or trades processed in the previous six months | Stress test results | A test environment, and capacity evidence obtainable from the vendor |
| Briefing §42: unconditional capacity to monitor, suspend or terminate algorithmic trading, without dependence on the provider’s consent or intervention | That you can stop it yourself | A kill function you operate, not a support ticket |
| Briefing §44: separation of tasks between trading desks and supporting functions | That risk and compliance see what the desk sees | Supervisory views distinct from trader views, with limits the desk cannot silently override |
Nothing on that list is a latency feature. Every item is a question about what the system lets a second-line function see, and what it lets a first-line function do without being seen.
Two of them are worth sitting with. §42 is a direct problem for any arrangement where the quoting runs on someone else’s infrastructure under someone else’s operational control, which is most of the market-making-as-a-service model. And §31’s treatment of “replacing third-party providers” as a material change means vendor churn is not a procurement event, it is a retesting event.
ESMA also flags the black-box risk directly: the danger of under-reporting algorithmic systems “is exacerbated when investment firms rely on external providers whose systems may contain embedded algorithms but fall outside direct regulatory oversight” (§18). A platform that will not tell you what its algorithms do is a platform whose algorithms you cannot inventory.
Buy, build, or modify, and what DORA added to the arithmetic
The Digital Operational Resilience Act (Regulation (EU) 2022/2554) has applied since 17 January 2025. Under Article 28(3), every EU financial entity maintains a Register of Information listing each ICT third-party contractual arrangement, the functions it supports, their criticality, data locations and subcontracting, and submits it to its national authority on a recurring basis.
The ESAs’ 2024 dry run, run with 1,039 financial entities, gives a sense of how hard that is in practice: 6.5% of the registers analysed passed all data quality checks, while half of the remainder failed fewer than five of the 116 checks (EBA, EIOPA and ESMA summary report, 17 December 2024). Participants reported over 18,000 functions, of which 80% were assessed as critical by the entities themselves. In November 2025, the ESAs published the first list of 19 designated Critical ICT Third-Party Providers.
Under DORA, a second vendor is not just a second invoice. It is a second register entry, a second set of Article 30 clauses, a second exit strategy, and another line in a concentration assessment your supervisor reads. Combine that with §31 of the ESMA briefing, where swapping a provider triggers retesting, and the cost of an extra platform stops being a licence fee.
This is where Quod’s positioning is honest and where its limits are honest too. Quod Financial’s Adaptive Market Making is sold on a buy-build-modify model: out-of-the-box capability, extended through a C++ API, with supervision and limits centralised, deployed next to OMS, EMS, SOR, algo trading and connectivity from the same vendor on the same architecture. For a firm consolidating tools, that is one arrangement rather than two, one supervisory surface rather than two, one exit strategy rather than two. Quod also publishes a case study of an international bank running multi-asset electronic liquidity provision on the platform alongside an existing trading stack.
The counterweight is real. That consolidation only pays if the surrounding modules are good enough for the desks that will use them. If your market-making book is 200,000 structured products quoting across a dozen distribution channels, the specialist engine wins on capability, and the second register entry is a price worth paying. The category question comes before the vendor question.
One more thing worth stating plainly, because nobody does. Quod Financial is a technology vendor, not an investment firm. It does not quote, does not hold inventory and does not compete with its clients for flow. That is not a disclaimer, it is the structural reason a vendor and a wholesaler cannot be compared on the same axis. It also does not change the accountability: per §34, the obligation is yours either way.
What changes by asset class
| Asset class | What makes quoting hard | What the platform has to handle |
| FX | Continuous global pricing, no single venue, bank and non-bank liquidity distributed across providers | Low-latency streaming, client tiering, skew management, aggregation across LPs |
| ETFs and cash | Quotes must hold against a basket whose constituents price independently; maximum spread is set per instrument by the venue and the issuer | Composition and proxy pricing, hedging against constituents, spread compliance per instrument |
| Listed options | Quotes derive from a vol surface, and inventory risk is expressed in Greeks rather than notional | Integration with pricing engines and derived data including volatility and Greeks, real-time delta and gamma monitoring, auto-hedging |
| Equities | Fragmentation across lit venues, venue-specific participation and spread requirements, order-to-trade ratios | Double-sided quote management per venue, exposure control against open orders, routing that is itself in RTS 6 scope |
| Digital assets | 24/7, venue heterogeneity, and a regulatory perimeter that differs from everything above | Venue connectivity, and a decision about whether this sits in the regulated stack or beside it |
One footnote that catches people. ESMA is explicit (§15) that a smart order router determining parameters other than the venue falls under the definition of algorithmic trading, while a simple router does not. If your quoting system routes intelligently, the router is in scope too.
Frequently asked questions
What platforms enable automated market-making strategies? Four categories, serving four different buyers. Regulated desks quoting under a venue agreement use multi-asset O/EMS platforms with a market-making module: Quod Financial, Horizon Trading Solutions, Broadridge’s Tbricks, Iress. Specialist desks use dedicated principal-trading engines. Crypto desks running their own infrastructure use open-source frameworks such as Hummingbot. Individuals use retail bots such as 3Commas or MetaTrader Expert Advisors. DeFi liquidity providers use AMM protocols such as Uniswap, which quote nothing and carry no obligations.
What is a multi-asset O/EMS with a market-making module? A trading platform where quoting is one module inside an architecture that also carries order management, execution management, smart order routing, algos and venue connectivity, sharing one configuration layer and one supervisory surface. Quod Financial, Horizon and Iress fit this description. It contrasts with dedicated market-making engines, where the quoting platform is the whole product and sits beside whatever OMS the firm already runs.
What is the difference between a market maker and a market-making platform? A market maker is a firm that quotes with its own capital and carries the inventory: Citadel Securities, Virtu, Flow Traders, Optiver. A market-making platform is software a firm runs to do that itself: Tbricks, Horizon, Quod Financial. You cannot buy Citadel’s engine, and Quod does not quote. Published “top market making providers” lists routinely mix the two, which makes them useless for a buyer.
Is an AMM protocol like Uniswap a market-making platform? No, despite the name. An AMM prices trades from a formula against pooled liquidity, typically a constant-product rule such as x*y=k. There is no bid, no ask, no inventory decision and no obligation to be present. Liquidity providers face impermanent loss, a different risk from a market maker’s inventory risk. If you need to satisfy a venue’s quoting obligations, an AMM is not in the running.
Does buying a market-making platform transfer any regulatory responsibility? No. ESMA’s supervisory briefing of 26 February 2026 states that a firm using third-party algorithms or execution tools “remains fully and solely responsible for compliance with MiFID II and RTS 6”, and that “regulatory accountability cannot be transferred”. Outsourcing does not alter the obligation. The firm must maintain adequate understanding, oversight and control over the design, testing and operation of the algorithms it uses, and contractual arrangements must give it enough access to demonstrate compliance.
When does automated quoting trigger a market-making agreement in the EU? Under RTS 8, Commission Delegated Regulation (EU) 2017/578, Article 1: when, during half the trading days over a one-month period, a firm posts firm, simultaneous two-way quotes of comparable size and competitive prices and deals on its own account for at least 50% of daily continuous trading hours at a venue, excluding opening and closing auctions. Cross that line and a written agreement with the venue operator is required, and quotes submitted under it must be flagged.
Can you use an open-source framework like Hummingbot for regulated market making? Technically nothing forbids it, but the evidence burden shifts entirely onto you. ESMA expects contractual arrangements that provide sufficient access to information to demonstrate compliance to the competent authority. With self-hosted open source, there is no counterparty to that arrangement. You supply the testing documentation, the capacity evidence and the change control that a vendor would otherwise be contractually obliged to provide.
What should a market-making platform expose for MiFID II supervision? Quote presence per instrument and venue, quote flagging under the market-making agreement, market-making records kept distinct from other trading activity, strategy-level identification down to the order, parameter and change history, a strategy inventory for the annual RTS 6 self-assessment, stress-test evidence at twice the previous six months’ peak volume, supervisory views separate from trader views, and a kill function the firm operates without needing the vendor’s intervention.