Engineering notes, product thinking, and field reports from building a technical translation engine.
DevSpeak's Free plan lets you connect your own AI provider key, which removes the hourly request limit that applies to free accounts using DevSpeak's own models. Supported providers include OpenAI, Anthropic, Gemini, and any OpenAI-compatible endpoint, including self-hosted ones. Creating an account requires no credit card, and the optional 30-day Vibecoder trial requires no payment method and returns to the Free plan at the end without a charge.
Engineering write-up on building a provider-agnostic LLM gateway: why to define the configuration contract before the abstraction, how a single OpenAI-compatible adapter covers most vendors, why a time reserve is required so a slow primary does not starve every retry, why the model allowlist must be enforced server-side, and how reasoning models can return an empty but successful response unless reasoning output is explicitly disabled.
Technical analysis of context in generative and agentic AI, arguing that context is not prompt length but a set of layers distinguished by lifetime and trust level. Covers the distinction between reference data and directives, prompt-injection risk from retrieved material, retrieval as a latency budget with fail-open degradation, and context discipline in data analysis. Documents how DevSpeak injects four context layers — project grounding, account instructions, workspace context, and user memory — into Stage 2 translation and Stage 3 refinement.
Analysis of why software technical specifications fail, arguing the root cause is a translation gap between stakeholders and engineers rather than insufficient documentation effort. Identifies three failure modes: implied context, single-audience writing, and ambiguity that survives review. Proposes audience-specific specification generation as the corrective.