[{"data":1,"prerenderedAt":327},["ShallowReactive",2],{"service-ai-integrations":3,"service-ai-integrations-related":123},{"id":4,"title":5,"body":6,"description":33,"extension":39,"eyebrow":40,"faq":41,"fromPrice":60,"headline":61,"includes":62,"meta":69,"navigation":70,"order":71,"path":72,"pillars":73,"priceNote":86,"related":87,"scales":90,"seo":96,"seoDescription":97,"seoTitle":98,"slug":99,"stack":100,"stem":119,"subhead":120,"tier":121,"__hash__":122},"services\u002Fservices\u002Fai-integrations.md","AI integrations",{"type":7,"value":8,"toc":32},"minimark",[9,14,18,22,25,29],[10,11,13],"h2",{"id":12},"what-a-starter-integration-looks-like","What a starter integration looks like",[15,16,17],"p",{},"The $18k floor covers a single AI feature inside your app, with prompts, structured outputs, cost ceilings, an eval suite, and an observability dashboard. Two to four weeks. The point is the smallest feature that earns its keep, not a sweeping AI rebrand of your product.",[10,19,21],{"id":20},"where-the-build-scales","Where the build scales",[15,23,24],{},"Adding features. Each new AI feature shares the eval harness and observability layer but has its own prompt design, output schema, and edge cases. Embedding pipelines over live data are a separate chapter (indexing, freshness, retrieval quality). Voice and vision change the cost and latency profile entirely.",[10,26,28],{"id":27},"what-we-will-not-build","What we will not build",[15,30,31],{},"AI features whose job is to write your marketing copy in your own voice (use Claude directly, save the integration cost). AI features with no eval. Anything where the success criteria is \"feels smart\" instead of a measurable lift.",{"title":33,"searchDepth":34,"depth":34,"links":35},"",2,[36,37,38],{"id":12,"depth":34,"text":13},{"id":20,"depth":34,"text":21},{"id":27,"depth":34,"text":28},"md","LLM FEATURES IN YOUR PRODUCT",[42,45,48,51,54,57],{"q":43,"a":44},"How is this different from AI agents?","AI agents are a build, the whole product is the agent. AI integrations are a feature added to a product that already exists. Different scope, different floor price.",{"q":46,"a":47},"We already use ChatGPT internally. Why this?","That is sometimes the right answer. We tell you when it is. When the workflow needs to run inside your app on customer data with audit and cost control, an integration is the next step.",{"q":49,"a":50},"How quickly do you ship?","Two to four weeks from kickoff for a single integration at the $18k floor. Multi-feature integration work runs four to eight.",{"q":52,"a":53},"What if our data is sensitive?","Zero-retention modes on Claude and OpenAI by default. Self-hosted vector stores where compliance requires it. We design for the strictest data class in the system.",{"q":55,"a":56},"Can you work in our existing codebase?","Yes. We are comfortable in Next, Nuxt, Rails, Django, Laravel, and most TypeScript or Python stacks. We will ask for a read-only access window during discovery.",{"q":58,"a":59},"Do you maintain it after launch?","Two-week support window included. Beyond that we offer monthly retainers for model-update and eval-regression work, sized to the integration.","$18k","Add AI to the app you already have.",[63,64,65,66,67,68],"One AI feature wired into your existing app","Prompt scaffolding, structured-output schemas","Token budget and cost-ceiling enforcement","Eval suite with 20+ test cases against your data","Observability dashboard for token spend per session","Two-week support window after launch",{},true,40,"\u002Fservices\u002Fai-integrations",[74,77,80,83],{"title":75,"body":76},"Real data, real evals","We test against your data, not synthetic prompts. Eval suite ships with the integration so you can detect regressions on model swaps.",{"title":78,"body":79},"Cost-aware by default","Token budgets per request, cheapest-model-first routing, caching where the input shape allows. The bill stays predictable.",{"title":81,"body":82},"Inside your codebase","We add the feature to the app you have. No separate \"AI service\" your team has to maintain in parallel. The seams live with the rest of the code.",{"title":84,"body":85},"Owned by you","Prompts, evals, API keys, observability all in your accounts. We don't run anything for you after handoff.","Phase-by-phase quote, fixed bid",[88,89],"ai-agents","saas",[91,92,93,94,95],"Multiple AI features sharing a common eval harness","Embedding pipelines over large or live datasets","Voice or vision integrations (image generation, OCR, transcripts)","Fine-tuning or distillation onto cheaper models","Customer-facing chat with retrieval and tools (see AI agents)",{"title":5,"description":33},"Add AI features to an existing product. Smart search, summarization, generation, structured extraction. Hand-coded against your real data with cost ceilings and evals. Starting at $18k.","AI Integration Services | LLM Features for Your SaaS | HARTECHO","ai-integrations",[101,104,107,110,113,116],{"name":102,"note":103},"Claude Opus \u002F Sonnet \u002F Haiku","Default. Routed by step based on cost and latency.",{"name":105,"note":106},"OpenAI \u002F Gemini","When the work calls for a second provider or a router pattern.",{"name":108,"note":109},"Postgres + pgvector","Embeddings, retrieval, semantic search in the database you already have.",{"name":111,"note":112},"Vercel AI SDK or direct Anthropic SDK","Whichever fits your existing app surface.",{"name":114,"note":115},"Langfuse \u002F Helicone","Trace every call, eval every output.",{"name":117,"note":118},"Inngest \u002F BullMQ","Background jobs for batch or long-running calls.","services\u002Fai-integrations","Smart search, summarization, generation, embedding-based recommendation, structured extraction. Wired into your existing codebase, evaluated against your real data.","core","-j0wBi04IhfAdX3TDzv1ND14cBJTY_C0VD3EccwE4CI",[124,223],{"id":125,"title":126,"body":127,"description":33,"extension":39,"eyebrow":151,"faq":152,"fromPrice":171,"headline":172,"includes":173,"meta":179,"navigation":70,"order":180,"path":181,"pillars":182,"priceNote":86,"related":194,"scales":195,"seo":201,"seoDescription":202,"seoTitle":203,"slug":88,"stack":204,"stem":220,"subhead":221,"tier":121,"__hash__":222},"services\u002Fservices\u002Fai-agents.md","AI agents",{"type":7,"value":128,"toc":146},[129,133,136,138,141,143],[10,130,132],{"id":131},"what-a-starter-ai-agent-looks-like","What a starter AI agent looks like",[15,134,135],{},"The $25k floor covers a single agent with up to five tools, RAG over one document corpus, an eval suite, cost-ceiling enforcement, and a real deployment. Two to four weeks. We pick the smallest agent that earns its keep and ship that one first.",[10,137,21],{"id":20},[15,139,140],{},"Three things drive the number up. The first is the number of agents and how they coordinate, which we charge against because multi-agent eval surfaces are exponential, not linear. The second is data scope, since RAG over five corpora and live data is a different problem than RAG over one. The third is the I\u002FO surface, since voice or webhooks or document parsing each add their own eval and reliability work.",[10,142,28],{"id":27},[15,144,145],{},"Agents that are just GPT wrappers with no tools. Agents whose only job is to summarize text. Agents with no eval set. Anything where the demo works on day one and the cost trace on day thirty is unbounded.",{"title":33,"searchDepth":34,"depth":34,"links":147},[148,149,150],{"id":131,"depth":34,"text":132},{"id":20,"depth":34,"text":21},{"id":27,"depth":34,"text":28},"AI AGENT DEVELOPMENT",[153,156,159,162,165,168],{"q":154,"a":155},"Why pick HARTECHO over hiring directly?","A two-week proof on a fixed bid usually answers \"is this even feasible\" faster than a hire-and-onboard cycle. Once it works we can hand it off to whoever you bring in.",{"q":157,"a":158},"Will the agent leak our data?","We default to a self-hosted vector store, prompt-injection filters, and no training on your data. Anthropic and OpenAI both offer zero-retention modes which we wire up by default.",{"q":160,"a":161},"What if Claude or OpenAI prices change?","Cost ceilings live in the codebase, not in the prompt. We can swap models or providers without rewriting the agent. The eval suite tells us whether the swap regressed quality.",{"q":163,"a":164},"How long does a starter agent take to ship?","Two to four weeks from kickoff to a deployed agent at the $25k floor. Multi-agent orchestration and voice push that to six to ten weeks.",{"q":166,"a":167},"Do you build chatbots?","Yes, but only when a chatbot is the right shape. Most \"we need a chatbot\" briefs are actually agent-with-tools briefs, which is a different build.",{"q":169,"a":170},"Can the agent run inside our existing app?","Yes. We package as an API the rest of your app calls. We have no preference about your frontend.","$25k","Custom AI agents built for production.",[174,175,176,177,178,68],"One agent with up to five tools","RAG over a single corpus, up to 5k documents","Eval suite with 30+ test cases","Cost-ceiling enforcement and structured fallbacks","Deployed to Vercel, Cloudflare, or your own infra",{},10,"\u002Fservices\u002Fai-agents",[183,186,189,192],{"title":184,"body":185},"Built for production, not demos","Eval suites before launch, cost ceilings per request, structured fallbacks when the model misses. The agent does not surprise you on the bill or the output.",{"title":187,"body":188},"Real tools, real side effects","Tool calling against your own APIs. We wire the agent into the systems that move money or change state, not just a chat with retrieval.",{"title":190,"body":191},"Your data, your boundaries","RAG over the documents you control. Citations on every answer. The model sees what you let it see and nothing else.",{"title":84,"body":193},"Codebase, prompts, eval set, deployment all hand off at launch. No vendor lock to a proprietary agent platform.",[99,89],[196,197,198,199,200],"Multi-agent orchestration with handoffs and shared memory","RAG across multiple corpora or live data sources","Voice agents (Cartesia, ElevenLabs) on top of the text core","Custom fine-tuning or distillation onto smaller models","Continuous-eval pipeline with regression alerts",{"title":126,"description":33},"Custom AI agent development for production. Autonomous workflows, RAG systems, LLM-powered chat, agentic platforms. Built on the Claude Agent SDK with real evals and cost ceilings. Starting at $25k.","AI Agent Development | Custom AI Agents Built for Production | HARTECHO",[205,208,211,213,215,218],{"name":206,"note":207},"Claude Agent SDK","Anthropic's first-party agent runtime. Primary for new builds.",{"name":209,"note":210},"Claude Opus \u002F Sonnet","Model selection per step based on cost and latency budgets.",{"name":105,"note":212},"Where the work calls for a second provider or a model-router.",{"name":108,"note":214},"Embeddings, retrieval, citation source-of-truth.",{"name":216,"note":217},"Inngest \u002F Trigger.dev","Long-running agent jobs with retries and observability.",{"name":114,"note":219},"Trace every call, eval every output, see real cost per session.","services\u002Fai-agents","Autonomous workflows, RAG, LLM-powered chat, agent platforms. Hand-coded against the Claude Agent SDK and equivalent tooling, deployed against real cost ceilings and real eval suites.","v5wiaSx8rMyfTHxm47TJRNdOBfRbWb4Riem2ujEXyKk",{"id":224,"title":225,"body":226,"description":33,"extension":39,"eyebrow":250,"faq":251,"fromPrice":270,"headline":271,"includes":272,"meta":279,"navigation":70,"order":280,"path":281,"pillars":282,"priceNote":86,"related":294,"scales":296,"seo":302,"seoDescription":303,"seoTitle":304,"slug":89,"stack":305,"stem":324,"subhead":325,"tier":121,"__hash__":326},"services\u002Fservices\u002Fsaas.md","SaaS \u002F web app",{"type":7,"value":227,"toc":245},[228,232,235,237,240,242],[10,229,231],{"id":230},"what-a-starter-saas-build-looks-like","What a starter SaaS build looks like",[15,233,234],{},"The $60k floor covers a multi-tenant data model, auth with invites, Stripe billing, an internal admin scaffold, and one core feature loop fully built. Eight to twelve weeks. The point is to land at a working app one user could actually pay for, with the seams in place to grow from there.",[10,236,21],{"id":20},[15,238,239],{},"Multiple feature loops is the most common scaler. Each new loop usually touches the data model, the UI, billing, and admin, so the cost is not linear. Real-time collaboration is a step change in complexity. Public APIs and webhooks add their own contracts to maintain. Enterprise plans (SSO, SCIM, audit) are usually deferred to a separate phase.",[10,241,28],{"id":27},[15,243,244],{},"SaaS that bets the company on a feature no user has asked for. SaaS without billing. Apps where the founder cannot describe the first paying user. Anything where the build is the cheap part and the distribution is unsolved.",{"title":33,"searchDepth":34,"depth":34,"links":246},[247,248,249],{"id":230,"depth":34,"text":231},{"id":20,"depth":34,"text":21},{"id":27,"depth":34,"text":28},"SAAS DEVELOPMENT",[252,255,258,261,264,267],{"q":253,"a":254},"What stage of company is this for?","Founders going from spec to first paying customer, or teams whose v1 was built fast and now needs to actually work. We are not the right shop if you already have ten engineers shipping daily.",{"q":256,"a":257},"How do you split the build from the design?","Discovery and design phases come first. We do not write code against unsolved problems. Once the screens are signed off the build is mostly mechanical.",{"q":259,"a":260},"Can we hand off mid-build?","Yes. Every milestone is a working app. You can stop after design, after the first feature loop, or after launch and we hand off the codebase as-is.",{"q":262,"a":263},"Why Nuxt or Next instead of Rails or Django?","We are faster in TypeScript and the audience usually has TypeScript people on hand for the next hire. We will write Python or Rails for the right project but it is not our default.",{"q":265,"a":266},"How does enterprise stuff fit in?","SSO, SCIM, and audit logs are a real chapter, not a checkbox. We sequence them after product-market fit. Building enterprise-features-first is a common reason for SaaS projects to die.",{"q":268,"a":269},"What about AI in the app?","See AI integrations. Adding LLM features to an existing SaaS is its own scope. We will tell you when AI is the right shape and when it is not.","$60k","Multi-tenant apps, hand-coded.",[273,274,275,276,277,278],"Multi-tenant data model with org + user + roles","Auth (email + OAuth) and team invites","Stripe Billing with one to two plan tiers","Internal admin scaffolding for support and ops","Errors, traces, uptime monitoring at launch","One core feature loop (the thing the app actually does) end-to-end",{},30,"\u002Fservices\u002Fsaas",[283,286,289,292],{"title":284,"body":285},"Real auth, day one","Multi-tenant from the start. Roles, invites, audit trail. NextAuth or Clerk, with the seams in your codebase rather than a black box.",{"title":287,"body":288},"Billing that holds up","Stripe subscriptions, metered usage, proration, dunning. Tested against a real failure-case suite so you don't lose revenue to an edge case.",{"title":290,"body":291},"Admin from sprint one","An internal admin gets built alongside the app, not bolted on after support tickets start. You can see and edit every record in your data.",{"title":84,"body":293},"Schema, migrations, deploy keys, observability. All in your accounts. We work for you, the code does too.",[99,295],"internal-tool",[297,298,299,300,301],"Multiple feature loops with cross-team workflows","Real-time collaboration (presence, cursors, shared state)","Public API and webhook system for customer integrations","SSO, SCIM, audit logging for enterprise plans","Mobile companion app (see Mobile)",{"title":225,"description":33},"Custom SaaS and web app development. Multi-tenant architecture, Stripe billing, admin dashboards, integrations. Hand-coded in Nuxt or Next, owned by you. Starting at $60k.","Custom SaaS Development | Web App MVP and Beyond | HARTECHO",[306,309,312,315,318,321],{"name":307,"note":308},"Nuxt 3 or Next.js","SSR-first. Choice depends on your stack and your team.",{"name":310,"note":311},"TypeScript strict","Types derived from schema, no any escapes.",{"name":313,"note":314},"Drizzle or Prisma + Postgres","Migrations checked in, no manual schema drift.",{"name":316,"note":317},"NextAuth \u002F Clerk \u002F Better Auth","Multi-tenant patterns from day one.",{"name":319,"note":320},"Stripe Billing","Subscriptions, metered usage, customer portal.",{"name":322,"note":323},"Sentry, Datadog, Better Stack","Errors, traces, uptime. All set up at launch.","services\u002Fsaas","Web applications with real auth, real billing, real admin, real integrations. Built so the second hire on your team can actually read the codebase.","t8bDHNs9JzEHooUJw_GA9eJOhLDCTWQazFFWKtCJy-g",1779549900336]