AI assistants and search decision
Claude
Build a narrow Claude-like assistant (chat UI + API integration + context retrieval) yourself — it's realistic for a single developer using Anthropic’s API and open-source tooling; but you’ll lose the proprietary models, vetted integrations, and managed enterprise safety/scale that drive Claude’s durable value.
Visit website↗Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need — the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship.
What a replacement has to do
- Call Anthropic Claude API for chat/completions, manage context (conversation state + embeddings), render UI, store/retrieve short-term cache and user history, and add prompt templates and safety checks.
What it still won’t have
- Managed model updates and new model access cadence
- Enterprise-grade safety tuning and monitoring
- Built-in connectors (vetted integrations) and hosted Workbench
- BAA/HIPAA-ready managed configuration unless negotiated with Anthropic
What remains hard
- Proprietary models
- Integration maintenance
- Infrastructure at scale
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 2 seats.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a minimal Claude-like assistant using the Anthropic Claude API and a lightweight web UI (stack: Next.js + React frontend, Node.js/Express backend, PostgreSQL for history, and Redis for short-term cache). Core features in scope: API auth and request wrapper, conversation state management with truncation, prompt templates (system/user/instruction), embeddings-based doc retrieval (use Anthropic embeddings or local OpenAI/HuggingFace embeddings), a simple web chat UI, and basic safety/retry/error handling. Out of scope: training models, large-scale multi-tenant infra, enterprise SOC/BAA compliance, and managed connectors. Require error handling, unit tests for API wrapper and conversation truncation, and integration test for end-to-end chat flow.
How we checked
How the score was reached
- Pay verdict base20
- An open-source build was found+5
- 6 cited sources+3
- Hard moats found in the evidence-6
- Evidence score22
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time — so the same evidence always produces the same number.
How scoring works →Cited sources · 6
Every page the run actually retrieved.
- official productClaude (landing / product)
- official productClaude Sonnet 5 announcement (pricing)
- official docsClaude API docs / overview
- official docsClaude docs — connectors overview
- open sourceopenclaw/openclaw
- open sourcearc53/DocsGPT
Integrity checks
What held up, and what did not.
