The Capabilities Behind a Swappable LLM Interface
A hosted model accepted our request format but leaked reasoning markers into citation parsing. The case traces capabilities, stream errors and separate speech.
Technical reviewer: Victor Sineglazov, AI Scientific Advisor.
Our hosted model accepted the same request format as the local adapter. Its answer still changed the meaning of our citation output. Reasoning text arrived inside ordinary content, and the citation parser counted markers from that reasoning as if they belonged to the answer.
The September 21, 2026 spike reported six cited sources even though the answer itself cited only two. Some extra markers came from passages the model had discussed as irrelevant. Moving the request to another endpoint had worked. Preserving the answer contract required more than compatible JSON.
We were building a professional learning platform for aesthetic medicine. The LLM boundary created on September 16 let the backend choose model configurations without exposing provider details throughout the product. The later spike showed which parts of that boundary needed behavioral checks.
A model registry has to change what the caller is allowed to do
The internal port offered complete and streaming responses plus model capabilities. Its registry supplied the provider model name, context and output budgets, supported options and concurrency limits. The backend could identify which configured model had answered.
These fields were useful only when code enforced them. A context budget constrained the material supplied to the model. A sampling setting determined which fields the adapter sent. A capability flag prevented an unsupported input from being quietly discarded.
| Boundary | What needed an explicit rule | Failure to watch for |
|---|---|---|
| Input capabilities | Reject images when the selected model lacks vision support. | The caller receives an answer to an input that was silently reduced. |
| Stream content | Distinguish answer text from reasoning and terminal metadata. | Non-answer text changes visible output or citations. |
| Usage accounting | Normalize input, output and cached-input counts. | Different field names produce misleading internal metrics. |
| Failure after output begins | Preserve partial output and represent the terminal error. | A retry appends another generation to an already visible answer. |
| Speech input | Return transcript segments and their time coordinates. | A text-completion abstraction loses the information retrieval needs. |
This table describes our design responsibilities. It is not a claim that current provider APIs expose identical features or that a capability list measures answer quality.
The parser was reading a larger text than the reader's answer
During the hosted-model spike, a reasoning block appeared in the content field. The answer's source markers and the reasoning's source markers therefore reached the same citation parser. The parser could resolve a marker correctly while applying it to the wrong part of the output.
The recorded diagnosis called for separating reasoning at the serving boundary and protecting the backend's citation parsing. We treat that as the diagnosis from the spike, rather than claiming that an endpoint change alone fixed it or that every later configuration was rechecked for this publication.
The case sharpened an existing distinction. A parser checks marker syntax and resolution; it does not establish that a claim is supported by the cited text. Our fifteen-question citation case describes that second failure. Here, even the set of text eligible for parsing needed a contract.
Retry rules depend on whether output has escaped
The adapter design distinguished a failure before the first response byte from a failure after output had begun. Before output, an eligible transient failure could be retried under the configured policy. After output, the client already held part of a generation.
The latter path preserved the delivered deltas and carried a normalized terminal error. It did not restart generation as though nothing had happened. Provider addresses and raw failure details were kept out of the public error message.
This boundary belongs to the adapter contract, but it does not settle durable conversation state. The streaming persistence case covers the separate work of saving partial answers, claiming generation and ordering a terminal event after its database write.
Extend the common port only while its semantics still fit
The later image work added image-bearing messages and an explicit vision capability. A request with images was rejected before the provider call if the selected model lacked that capability. The desired failure was visible refusal of an unsupported operation, rather than silent loss of part of the request.
The configuration also separated the model answering questions from the model describing images during indexing. Choosing an answer model did not automatically authorize it to process every uploaded source image.
Speech took another path. Transcription used its own port and registry because the operation returned segments with time coordinates and had a job lifecycle. Pretending that a transcript was merely a chat response would have hidden information the ingestion pipeline needed.
The separation followed the operation's output and lifecycle. It did not imply that every new model capability deserved another service or another public API. For image input, extending the existing message contract was workable; for transcription, the result had a different shape.
Keep the implementation decision attached to its original scale
For the September implementation, we chose a small HTTP adapter behind the internal port. The decision record tied that choice to the small provider set and left room to reconsider it when additional native integrations were needed.
That historical decision is not evidence that today's model frameworks lack a feature, or that a custom adapter is cheaper at every integration count. We have not compared current framework releases in this article.
The reusable artifact is the contract test set: inspect the outbound fields, translate usage and stop information, test unsupported input, then test a failure before and after visible output. Run a real model path as well as deterministic doubles. The hosted reasoning failure passed through the same adapter shape that already looked interchangeable.
Evidence: the September 16 LLM-port specification, September 21 hosted-model spike and later image and speech updates in our development records. Provider behavior is described for those recorded configurations. No current API comparison, quality-equivalence study or fresh model benchmark is claimed.
FAQ
Quick answers to common questions about custom software development, pricing, process and technology.
Type to filter questions and answers. Use Topic to narrow the list.
Showing all 14
No matches
Try a different keyword, change the topic or clear filters
-
How long have you been in business?
Pharos Production has been in business since 2013, with over 13 years of experience in custom software development. During this time, we have delivered over 110 applications for 200+ clients across 28 industries, including FinTech, healthcare, crypto and e-commerce. We are rated 5/5 on Clutch based on 107 verified reviews (2026).
-
What services do you provide?
Pharos Production provides six core service categories: AI Development (LLM applications, AI agents, machine learning and computer vision), Blockchain and Web3 (smart contracts, DeFi, tokenization on Ethereum, Solana, TON and other chains), FinTech Development (payment platforms, digital banking, KYC and AML integration), Software Development (mobile apps, web platforms, database design, UI/UX), Security and QA (code audits, penetration testing, smart contract audits, manual and automated testing) and Consulting (architecture design, MVP validation, startup consulting).
-
Where is Pharos Production located?
Pharos Production is headquartered in Las Vegas, Nevada, USA (5348 Vegas Dr, Las Vegas, NV 89108), with an engineering office in Kyiv, Ukraine (44-B Eugene Konovalets Str. Suite 201, Kyiv 01133). We work with clients worldwide and provide remote collaboration across all time zones. Visit our contact page for directions and scheduling options.
-
How big is your team?
Pharos Production has a team of 90+ engineers, including software developers, blockchain specialists, QA engineers, DevOps experts, UI/UX designers, project managers and solution architects. Our founder, Dr. Dmytro Nasyrov, holds a PhD in Artificial Intelligence and leads the technical direction of all projects.
-
What types of customers do you serve?
We serve a wide range of clients, from startups and product companies to mid-sized enterprises and large institutions. Our clients include crypto exchanges, FinTech providers (like Pleenk), healthcare organizations, sportsbook operators (like Pro Gambling), e-commerce platforms and SaaS companies. Pharos Production has worked with 200+ clients across 28 industries since 2013, adapting engagement models to match each client's stage, whether it is MVP validation for a startup or enterprise-scale development for an established business.
-
How is a custom software development company different from an agency?
A custom software development company like Pharos Production assigns one dedicated engineering team that owns your product across multiple release cycles, instead of handing it off once a contract ends. Agencies more often rotate staff between clients and optimize for fast delivery over long-term product ownership.
Pharos Production keeps the same team on your codebase through 2-week sprints, so there is no re-onboarding cost each time you request a new feature.
-
How much does custom software development cost?
Custom software development costs vary based on project scope and complexity. At Pharos Production, typical project ranges are: MVP development ($10,000-$25,000), suitable for startups validating a product idea; full-fledged production ($25,000-$50,000), for established businesses scaling a proven concept; and full-cycle development ($50,000-$80,000+), for complex enterprise-grade systems. These ranges include architecture design, development, QA testing and deployment. Hourly rates range from $50 to $99 depending on role and seniority. Final pricing depends on technology stack, number of integrations and engagement model (staff augmentation, dedicated team or project outsourcing).
-
How long does it take to develop a custom software application?
Development timelines depend on scope and complexity. At Pharos Production, a typical MVP takes 2-4 months, a production-ready application takes 4-8 months and a complex enterprise system can take 8-12+ months. We use an agile methodology with 2-week sprints, delivering working increments after each sprint. Every sprint includes a retrospective, progress report and planning session for the next iteration. This approach ensures transparency and allows businesses to launch faster by prioritizing high-impact features first. Get a timeline estimate for your project.
-
What industries does Pharos Production serve?
Pharos Production serves 28 industries: FinTech, MiCA Compliance, Banking, Insurance, Crypto, Healthcare, Fitness, E-Commerce, Food & Restaurant, On-Demand, Manufacturing, Automotive, Energy, Oil & Gas, Supply Chain, Construction, Real Estate, Gaming, Streaming, Media, Telecom, Education, Aviation, Travel, Logistics, Marketing, Agriculture and Recommender Systems. Our deepest expertise is in FinTech, blockchain and healthcare, where we have delivered compliance-ready platforms (HIPAA, PCI DSS, GDPR) and high-load systems handling thousands of concurrent users. For the latest industry insights, read our guides on FinTech trends in 2026 and the Web3 technology stack.
-
Why hire a software development company instead of building in-house?
Hiring a software development company offers faster time-to-market, lower upfront costs and access to specialized expertise without long-term employment commitments. According to Deloitte's 2024 Global Outsourcing Survey, 57% of companies outsource software development to access skills they cannot hire internally.
Factor In-house team Software development company Time to assemble 3-6 months (recruiting + onboarding) 1-2 weeks Upfront cost High (salaries, benefits, equipment) Lower (project-based pricing) Specialized expertise Limited to who you can hire locally Access to 90+ engineers across blockchain, AI, FinTech Scalability Slow (each new hire takes months) Fast (scale up or down per sprint) Long-term commitment Full-time employment contracts Flexible engagement models Risk High if key engineers leave Company ensures continuity and knowledge transfer For businesses that need blockchain, AI or high-load architecture expertise, outsourcing to a specialized firm like Pharos Production reduces risk and accelerates delivery.
-
What projects does Pharos Production not take on?
Pharos Production focuses on mid-to-large custom software projects with budgets starting at $10,000. We do not take on template-based websites, WordPress theme customization or short-term contracts under one month. We also do not provide non-technical staffing (marketing, sales or design-only roles). Our strongest fit is blockchain, FinTech and healthcare projects where security, compliance and high-load architecture are critical. For smaller projects or MVPs under $10,000, we recommend exploring freelance platforms or no-code tools as a more cost-effective starting point.
-
Why does Pharos Production use agile methodology?
We use agile with 2-week sprints because it reduces the risk of building features that miss the mark. Each sprint ends with a working demo, a retrospective and a plan for the next iteration.
This means clients see progress every 14 days and can adjust priorities based on real results, not assumptions. We chose this approach after years of experience showing that rigid, fixed-scope contracts lead to scope creep, missed deadlines and products that do not match market needs by launch day.
-
When should you NOT hire a custom software development company?
Custom development is not the right choice in every situation. You should not hire a custom software company if: your problem is fully solved by an existing SaaS product (e.g. Shopify for e-commerce, Salesforce for CRM); your budget is under $10,000 and timeline is under 4 weeks; you need a simple landing page or marketing website (WordPress or Webflow is faster and cheaper); or you are still validating the idea and have not spoken to potential users yet.
In these cases, off-the-shelf tools or no-code platforms offer better ROI. Custom development makes sense when you need unique workflows, regulatory compliance, high-load architecture or competitive differentiation that packaged software cannot provide.
-
What kind of projects has Pharos Production delivered?
Here are three anonymized examples from our recent delivery history:
FinTech startup - payment platform (MVP)
Scope: mobile app + backend API with bank-grade encryption. Team: 4 engineers, 1 QA. Timeline: 10 weeks. Budget: $38,000. Result: launched on schedule, processed $2M+ in transactions within the first quarter.Healthcare provider - patient portal (Full product)
Scope: HIPAA-aligned web platform with EHR integration, appointment scheduling and telemedicine. Team: 6 engineers, 1 DevOps, 2 QA. Timeline: 6 months. Budget: $120,000. Result: 15,000+ active patients, zero compliance violations in two annual audits.Crypto exchange - trading engine (Complex)
Scope: high-load matching engine handling 50,000+ orders per second, multi-chain wallet infrastructure on Ethereum and Solana. Team: 8 engineers, 2 QA, 1 security auditor. Timeline: 11 months. Budget: $340,000. Result: 99.97% uptime, passed three independent security audits.See more projects: NoMoreBets, Pulse, Sagas, Gambit Stream and Pleenk. For the full portfolio, visit our case studies. Learn more about the technology behind these projects in our guide to stablecoins and crypto infrastructure.
Dmytro Ihnatov is the Lead Backend Developer at Pharos Production. He documents backend architecture and research experiments from the team's work on a professional learning platform for aesthetic medicine.
