https://kmail.at/tutorials Terminal lessons, technology deep-dives and field notes by Mohammed Kmail. en https://kmail.at/knowledge/field-notes/ai-regulatory-compliance https://kmail.at/knowledge/field-notes/ai-regulatory-compliance Sun, 16 Aug 2026 12:36:34 GMT KI-VO · AI Act · reallabore https://kmail.at/knowledge/field-notes/devops-reverse-proxy https://kmail.at/knowledge/field-notes/devops-reverse-proxy Sun, 16 Aug 2026 12:36:34 GMT traefik · nginx · docker https://kmail.at/knowledge/field-notes/data-analysis-notebooks https://kmail.at/knowledge/field-notes/data-analysis-notebooks Sun, 16 Aug 2026 12:36:34 GMT marimo · pandas · reports https://kmail.at/knowledge/field-notes/custom-llm-environments https://kmail.at/knowledge/field-notes/custom-llm-environments Sun, 16 Aug 2026 12:36:34 GMT open-webui · rag · ollama https://kmail.at/tutorials/rm https://kmail.at/tutorials/rm Sun, 16 Aug 2026 00:00:00 GMT Delete files and directories. Simple, permanent, and unforgiving — the command that demands respect. https://kmail.at/tutorials/docling https://kmail.at/tutorials/docling Sat, 15 Aug 2026 00:00:00 GMT Convert PDFs, DOCX, PPTX, images and web pages into clean Markdown, HTML or JSON with layout awareness, table recognition, OCR and provenance. https://kmail.at/tutorials/mv https://kmail.at/tutorials/mv Sat, 15 Aug 2026 00:00:00 GMT Relocate files between directories or give them a new name. Also the safe way to overwrite. https://kmail.at/tutorials/graphiti https://kmail.at/tutorials/graphiti Fri, 14 Aug 2026 00:00:00 GMT Build knowledge graphs that evolve over time, track what was true and when, and give AI agents rich, queryable context instead of flat document chunks. https://kmail.at/tutorials/cp https://kmail.at/tutorials/cp Fri, 14 Aug 2026 00:00:00 GMT Duplicate files and directories, preserving or changing their contents. The everyday copy tool. https://kmail.at/tutorials/lc-2026-08-14-langsmith-provider https://kmail.at/tutorials/lc-2026-08-14-langsmith-provider Fri, 14 Aug 2026 00:00:00 GMT langchain 1.3.15 adds LangSmith as a first-class provider to init_chat_model, so you can call LangSmith-hosted models through the same unified factory. https://kmail.at/tutorials/uv https://kmail.at/tutorials/uv Thu, 13 Aug 2026 00:00:00 GMT A single Rust binary that replaces pip, pip-tools, virtualenv, pyenv, and Poetry — with a lockfile, project management, and 10-100x faster installs. https://kmail.at/tutorials/chmod https://kmail.at/tutorials/chmod Thu, 13 Aug 2026 00:00:00 GMT Control who can read, write, and execute a file. The key to sane, secure permissions. https://kmail.at/tutorials/lc-2026-08-13-reasoning-effort-init https://kmail.at/tutorials/lc-2026-08-13-reasoning-effort-init Thu, 13 Aug 2026 00:00:00 GMT A provider-agnostic reasoning_effort parameter usable through init_chat_model, enabling cheap classification passes and expensive reasoning only where it pays. https://kmail.at/tutorials/ls https://kmail.at/tutorials/ls Wed, 12 Aug 2026 00:00:00 GMT The first command you learn. List files and directories with rich formatting. https://kmail.at/tutorials/grep https://kmail.at/tutorials/grep Wed, 12 Aug 2026 00:00:00 GMT The pattern-matching powerhouse. Search files and streams with regular expressions. https://kmail.at/tutorials/find https://kmail.at/tutorials/find Wed, 12 Aug 2026 00:00:00 GMT Search the filesystem by name, type, size, and time. More powerful than you think. https://kmail.at/tutorials/tar https://kmail.at/tutorials/tar Wed, 12 Aug 2026 00:00:00 GMT Bundle files into a single archive, optionally compressed. The universal backup and transfer tool. https://kmail.at/tutorials/marimo https://kmail.at/tutorials/marimo Wed, 12 Aug 2026 00:00:00 GMT A reactive notebook and app framework that runs cells on input change, stores notebooks as pure Python, and deploys as scripts or web apps. https://kmail.at/tutorials/gzip https://kmail.at/tutorials/gzip Wed, 12 Aug 2026 00:00:00 GMT Shrink a file with LZ77 + Huffman coding. The default compressor Linux ships with. https://kmail.at/tutorials/lc-2026-08-12-reasoning-effort-standard https://kmail.at/tutorials/lc-2026-08-12-reasoning-effort-standard Wed, 12 Aug 2026 00:00:00 GMT langchain-core 1.5.4 adds reasoning_effort as a standard, provider-agnostic chat model parameter, unifying how you control reasoning depth across providers. https://kmail.at/tutorials/lc-2026-08-11-openai-tool-call-filtering https://kmail.at/tutorials/lc-2026-08-11-openai-tool-call-filtering Tue, 11 Aug 2026 00:00:00 GMT langchain-openai 1.4.3 filters malformed tool calls from model content, so your agent loop only ever sees well-formed, callable entries. https://kmail.at/tutorials/lc-2026-08-10-openai-context-window-handling https://kmail.at/tutorials/lc-2026-08-10-openai-context-window-handling Mon, 10 Aug 2026 00:00:00 GMT langchain-openai 1.4.2 surfaces context overflow as a catchable openai.BadRequestError, enabling explicit truncation or model-switch recovery. https://kmail.at/tutorials/lc-2026-08-09-openai-context-window https://kmail.at/tutorials/lc-2026-08-09-openai-context-window Sun, 09 Aug 2026 00:00:00 GMT langchain-openai 1.4.2 wraps OpenAI context overflow as a proper LangChain exception you can catch, inspect, and recover from. https://kmail.at/tutorials/lc-2026-08-08-openai-context-window-error https://kmail.at/tutorials/lc-2026-08-08-openai-context-window-error Sat, 08 Aug 2026 00:00:00 GMT langchain-openai 1.4.2 catches ContextWindowExceededError and re-raises a clean error with the model name and token context. https://kmail.at/tutorials/lc-2026-08-07-anthropic-user-profile-id https://kmail.at/tutorials/lc-2026-08-07-anthropic-user-profile-id Fri, 07 Aug 2026 00:00:00 GMT langchain-anthropic adds a user_profile_id convenience attribute for tagging requests with a user profile for billing, analytics, and access control. https://kmail.at/tutorials/lc-2026-08-06-anthropic-tool-schema-composition https://kmail.at/tutorials/lc-2026-08-06-anthropic-tool-schema-composition Thu, 06 Aug 2026 00:00:00 GMT langchain-anthropic 1.5.4 catches unsupported top-level schema composition (oneOf, allOf, bare array) early instead of failing at runtime. https://kmail.at/tutorials/lc-2026-08-04-gateway-empty-env-vars https://kmail.at/tutorials/lc-2026-08-04-gateway-empty-env-vars Tue, 04 Aug 2026 00:00:00 GMT A deeper look at how langchain-core 1.5.2 distinguishes a variable that is unset (None) from one explicitly set to empty ( https://kmail.at/tutorials/lc-2026-08-03-gateway-empty-env https://kmail.at/tutorials/lc-2026-08-03-gateway-empty-env Mon, 03 Aug 2026 00:00:00 GMT langchain-core 1.5.2 treats an explicitly empty gateway env var as unset, falling back to LANGSMITH_API_KEY instead of failing auth silently. https://kmail.at/tutorials/lc-2026-08-02-gateway-key-fallback https://kmail.at/tutorials/lc-2026-08-02-gateway-key-fallback Sun, 02 Aug 2026 00:00:00 GMT The LangChain gateway now falls back to LANGSMITH_API_KEY when no explicit key is in config, aligning server-side init with the rest of the client. https://kmail.at/tutorials/langchain-langsmith-api-key-fallback https://kmail.at/tutorials/langchain-langsmith-api-key-fallback Sat, 01 Aug 2026 00:00:00 GMT langchain-core 1.5.3 lets the gateway fall back to LANGSMITH_API_KEY when LANGCHAIN_API_KEY is missing, ending manual sync. https://kmail.at/tutorials/langchain-gateway-empty-var-fallback https://kmail.at/tutorials/langchain-gateway-empty-var-fallback Fri, 31 Jul 2026 00:00:00 GMT langchain-core 1.5.2 falls back to the next credential source when a gateway env var is empty, ending silent auth failures. https://kmail.at/tutorials/langchain-empty-string-env-coerce https://kmail.at/tutorials/langchain-empty-string-env-coerce Thu, 30 Jul 2026 00:00:00 GMT langchain-core 1.5.2 coerces empty gateway env vars to None so LangServe and tracing stay stable in real deployments. https://kmail.at/tutorials/langchain-empty-gateway-env-crash https://kmail.at/tutorials/langchain-empty-gateway-env-crash Wed, 29 Jul 2026 00:00:00 GMT langchain-core 1.5.2 stops crashing when a gateway env var is set to an empty string instead of being unset. https://kmail.at/tutorials/langchain-gateway-api-key-env https://kmail.at/tutorials/langchain-gateway-api-key-env Tue, 28 Jul 2026 00:00:00 GMT LANGCHAIN_LANGCHAIN_GATEWAY_API_KEY lets one env var authenticate every LangChain model client against LangSmith Gateway. https://kmail.at/tutorials/langchain-tool-schema-cache-throughput https://kmail.at/tutorials/langchain-tool-schema-cache-throughput Mon, 27 Jul 2026 00:00:00 GMT langchain-core 1.5.1 refactors BaseTool token counting to serve the schema from a warm cache, cutting latency and token overhead. https://kmail.at/tutorials/langchain-token-counting-cache-schema https://kmail.at/tutorials/langchain-token-counting-cache-schema Sun, 26 Jul 2026 00:00:00 GMT count_tokens_approximately now leans on a per-tool tool_call_schema cache, making repeated counting far cheaper in high-frequency loops. https://kmail.at/tutorials/langchain-token-counting-determinism https://kmail.at/tutorials/langchain-token-counting-determinism Sat, 25 Jul 2026 00:00:00 GMT langchain-core 1.5.1 fixes count_tokens_approximately for BaseTool so identical tool schemas return stable counts every call. https://kmail.at/tutorials/langchain-tool-call-schema-cache https://kmail.at/tutorials/langchain-tool-call-schema-cache Fri, 24 Jul 2026 00:00:00 GMT langchain-core 1.5.1 memoizes each BaseTool tool_call_schema, slashing redundant serialization during high-frequency token counting. https://kmail.at/tutorials/langchain-reasoning-effort-standard https://kmail.at/tutorials/langchain-reasoning-effort-standard Thu, 23 Jul 2026 00:00:00 GMT LangChain unifies reasoning depth control across Grok, Claude, GPT-4o, and Fireworks with one standard chat-model parameter. https://kmail.at/tutorials/langchain-reasoning-effort-provider-knob https://kmail.at/tutorials/langchain-reasoning-effort-provider-knob Wed, 22 Jul 2026 00:00:00 GMT The langchain-core 1.5.0 release (PR #38887) normalises reasoning_effort into a single interface mapping to xai_reasoning_effort, thinking tokens, and more. https://kmail.at/tutorials/langchain-reasoning-effort-unified https://kmail.at/tutorials/langchain-reasoning-effort-unified Tue, 21 Jul 2026 00:00:00 GMT langchain-core 1.5.0 (PR #38887) introduces reasoning_effort as a standard parameter mapping to provider-specific thinking knobs across chat models. https://kmail.at/tutorials/langchain-meta-init-chat-model https://kmail.at/tutorials/langchain-meta-init-chat-model Mon, 20 Jul 2026 00:00:00 GMT langchain 1.3.13 (PR #38786) adds a meta extra so init_chat_model can spin up Meta-backed chat models through the unified ChatModel interface. https://kmail.at/tutorials/langchain-toolerror-middleware-callback https://kmail.at/tutorials/langchain-toolerror-middleware-callback Sun, 19 Jul 2026 00:00:00 GMT langchain 1.3.13 adds ToolErrorMiddleware with an on_error callback for logging, transforming, or falling back on tool failures at the infrastructure level. https://kmail.at/tutorials/langchain-toolerror-middleware-default https://kmail.at/tutorials/langchain-toolerror-middleware-default Sat, 18 Jul 2026 00:00:00 GMT langchain 1.3.14 (PR #38781) intercepts tool errors and wraps them in a structured response, optionally retrying or falling back to a default value. https://kmail.at/tutorials/langchain-toolretry-interrupt-revocation https://kmail.at/tutorials/langchain-toolretry-interrupt-revocation Fri, 17 Jul 2026 00:00:00 GMT The langchain 1.3.12 bug fix (PR #38722) stops ToolRetryMiddleware from swallowing cancel signals, fixing resource leaks and runaway loops. https://kmail.at/tutorials/langchain-openai-cache-config https://kmail.at/tutorials/langchain-openai-cache-config Thu, 16 Jul 2026 00:00:00 GMT langchain-openai 1.3.5 accepts a cache parameter plus CacheConfig so you control which message blocks stick in cache and for how long. https://kmail.at/tutorials/langchain-openai-cache-weight https://kmail.at/tutorials/langchain-openai-cache-weight Wed, 15 Jul 2026 00:00:00 GMT langchain-openai 1.3.5 accepts a cache_weight hint on ChatOpenAI so LangChain wires the right OpenAI caching structure for you. https://kmail.at/tutorials/langchain-toolretry-interrupt-propagate https://kmail.at/tutorials/langchain-toolretry-interrupt-propagate Tue, 14 Jul 2026 00:00:00 GMT langchain 1.3.12 (PR #38722) ensures an interrupt raised during a tool retry loop is honoured instead of being silently swallowed. https://kmail.at/tutorials/langchain-openai-cache-control-explicit https://kmail.at/tutorials/langchain-openai-cache-control-explicit Mon, 13 Jul 2026 00:00:00 GMT langchain-openai 1.3.5 lets you mark stable message blocks as cacheable with cache_control, so OpenAI caches them at the token level and discounts repeated context. https://kmail.at/tutorials/langchain-init-chat-model-meta https://kmail.at/tutorials/langchain-init-chat-model-meta Sun, 12 Jul 2026 00:00:00 GMT init_chat_model now supports a meta extra for instantiating langchain-meta backed chat models through one provider-agnostic interface. https://kmail.at/tutorials/langchain-openai-prompt-caching https://kmail.at/tutorials/langchain-openai-prompt-caching Sat, 11 Jul 2026 00:00:00 GMT langchain-openai 1.3.5 adds cache_implicit for explicit prompt caching, surfacing cache hits in additional_kwargs. https://kmail.at/tutorials/langchain-output-parser-bugfix https://kmail.at/tutorials/langchain-output-parser-bugfix Fri, 10 Jul 2026 00:00:00 GMT langchain-core 1.4.9 adds context-rich error messages and corrects XML continuation-line parsing for edge-case LLM output. https://kmail.at/tutorials/langchain-core-output-parsers https://kmail.at/tutorials/langchain-core-output-parsers Thu, 09 Jul 2026 00:00:00 GMT langchain-core 1.4.9 patches XML continuation-line parsing and surfaces clear ValidationError messages from Pydantic. https://kmail.at/tutorials/langchain-mistral-citation-metadata https://kmail.at/tutorials/langchain-mistral-citation-metadata Wed, 08 Jul 2026 00:00:00 GMT langchain-mistralai 1.1.6 promotes Mistral citations to structured AIMessage metadata for traceability and hallucination detection. https://kmail.at/tutorials/langchain-openrouter-default-headers https://kmail.at/tutorials/langchain-openrouter-default-headers Tue, 07 Jul 2026 00:00:00 GMT langchain-openrouter 0.2.6 adds a default_headers parameter to inject custom HTTP headers into every request. https://kmail.at/tutorials/langchain-mistral-citations-rag https://kmail.at/tutorials/langchain-mistral-citations-rag Mon, 06 Jul 2026 00:00:00 GMT langchain-mistralai 1.1.6 exposes Mistral citations as structured metadata on the AIMessage for grounded RAG answers. https://kmail.at/tutorials/langchain-anthropic-streaming-initial-text https://kmail.at/tutorials/langchain-anthropic-streaming-initial-text Sun, 05 Jul 2026 00:00:00 GMT langchain-anthropic 1.4.8 ensures the first streamed token is captured even when it shares a chunk with the block start marker. https://kmail.at/tutorials/langchain-provider-strategy-strict https://kmail.at/tutorials/langchain-provider-strategy-strict Sat, 04 Jul 2026 00:00:00 GMT langchain 1.3.11 makes bind_tools provider-aware, applying strict=True only where the provider actually supports it. https://kmail.at/tutorials/langchain-anthropic-content-block-start https://kmail.at/tutorials/langchain-anthropic-content-block-start Fri, 03 Jul 2026 00:00:00 GMT langchain-anthropic 1.4.8 stops dropping the initial text chunk that arrives with the content_block_start event. https://kmail.at/tutorials/langchain-parallel-tools-bind-vienna https://kmail.at/tutorials/langchain-parallel-tools-bind-vienna Thu, 02 Jul 2026 00:00:00 GMT Eliminate the round-trip tax with parallel_tool_calls via bind_tools on the OpenRouter integration, from two round-trips to one. https://kmail.at/tutorials/langchain-parallel-tools-bind-july https://kmail.at/tutorials/langchain-parallel-tools-bind-july Wed, 01 Jul 2026 00:00:00 GMT The parallel_tool_calls flag on bind_tools revisited in July, focused on fan-out/fan-in architectures and executing the returned calls. https://kmail.at/tutorials/langchain-parallel-tools-provider https://kmail.at/tutorials/langchain-parallel-tools-provider Thu, 25 Jun 2026 00:00:00 GMT The parallel_tool_calls knob plus the ProviderStrategy fix that stops strict=True from leaking to non-OpenAI providers. https://kmail.at/tutorials/langchain-parallel-tools-openrouter https://kmail.at/tutorials/langchain-parallel-tools-openrouter Wed, 24 Jun 2026 00:00:00 GMT Native parallel_tool_calls=True on the langchain-openrouter integration, with LangChain handling chunk reassembly automatically. https://kmail.at/tutorials/langchain-parallel-tools-bind https://kmail.at/tutorials/langchain-parallel-tools-bind Tue, 23 Jun 2026 00:00:00 GMT Enable parallel_tool_calls=True on bind_tools so your OpenRouter model fires multiple tools in a single response turn. https://kmail.at/tutorials/langchain-disallow-any-generics https://kmail.at/tutorials/langchain-disallow-any-generics Mon, 22 Jun 2026 00:00:00 GMT Enforce strict generic typing across LangChain components with the disallow_any_generics flag to catch type-safety bugs at the source. https://kmail.at/tutorials/langchain-streaming-usage-chunks https://kmail.at/tutorials/langchain-streaming-usage-chunks Sun, 21 Jun 2026 00:00:00 GMT Read cumulative token usage on each AIMessageChunk as it streams, enabling live cost meters and streaming-aware budgets. https://kmail.at/tutorials/langchain-tool-schema-memoization https://kmail.at/tutorials/langchain-tool-schema-memoization Sat, 20 Jun 2026 00:00:00 GMT BaseTool.tool_call_schema is now cached per tool instance in langchain-core 1.4.8 — a quiet performance win for tool-heavy chains. https://kmail.at/tutorials/langchain-preserve-usage-streaming-v3 https://kmail.at/tutorials/langchain-preserve-usage-streaming-v3 Fri, 19 Jun 2026 00:00:00 GMT Keep accurate token-usage metadata when you stream LLM responses via the v3 streaming protocol, instead of losing it to the ether. https://kmail.at/tutorials/lc-2026-06-18-normalize-streamed-tool-calls-fix https://kmail.at/tutorials/lc-2026-06-18-normalize-streamed-tool-calls-fix Thu, 18 Jun 2026 00:00:00 GMT langchain-openai 1.3.1 / core 1.4.7 standardize streamed tool-call events so AIMessageChunk.tool_calls is always a consistent, non-duplicated list of ToolCall objects. https://kmail.at/tutorials/lc-2026-06-17-normalize-streamed-tool-calls-agent https://kmail.at/tutorials/lc-2026-06-17-normalize-streamed-tool-calls-agent Wed, 17 Jun 2026 00:00:00 GMT langchain-openai 1.3.1 fixes the v1 stream parser so agentic pipelines get clean, complete ToolCallChunk objects with no duplicates or partial strings. https://kmail.at/tutorials/lc-2026-06-16-create-agent-overloads https://kmail.at/tutorials/lc-2026-06-16-create-agent-overloads Tue, 16 Jun 2026 00:00:00 GMT langchain 1.3.8 adds PEP 484 overloads to create_agent so type checkers infer the exact return type from your tools and model. https://kmail.at/tutorials/lc-2026-06-15-pydantic-v1-runnable https://kmail.at/tutorials/lc-2026-06-15-pydantic-v1-runnable Mon, 15 Jun 2026 00:00:00 GMT langchain-core 1.4.7 fixes serialization, deserialization, and tool-binding for Pydantic v1 schemas so legacy codebases get reliable tool calling. https://kmail.at/tutorials/lc-2026-06-14-basetool-pydantic-v1 https://kmail.at/tutorials/lc-2026-06-14-basetool-pydantic-v1 Sun, 14 Jun 2026 00:00:00 GMT langchain-core 1.4.7 fixes Pydantic v1 support in tools/runnable so custom tools build correct schemas on legacy Pydantic setups. https://kmail.at/tutorials/lc-2026-06-13-normalize-streamed-tool-calls-chunks https://kmail.at/tutorials/lc-2026-06-13-normalize-streamed-tool-calls-chunks Sat, 13 Jun 2026 00:00:00 GMT langchain-openai 1.3.1 and core 1.4.7 normalize streamed tool-call chunks so token-by-token output matches non-streamed ToolCall shape. https://kmail.at/tutorials/lc-2026-06-12-normalize-streamed-tool-calls https://kmail.at/tutorials/lc-2026-06-12-normalize-streamed-tool-calls Fri, 12 Jun 2026 00:00:00 GMT langchain-core 1.4.6 fixes how OpenAI v1 streamed tool call deltas are normalized into canonical ToolCall format, killing duplicates and dropped args. https://kmail.at/tutorials/lc-2026-06-11-provider-tool-search https://kmail.at/tutorials/lc-2026-06-11-provider-tool-search Thu, 11 Jun 2026 00:00:00 GMT langchain 1.3.7 introduces ProviderToolSearchMiddleware, which routes each query through a provider-native tool search step before the model responds. https://kmail.at/tutorials/lc-2026-06-10-summarization-middleware https://kmail.at/tutorials/lc-2026-06-10-summarization-middleware Wed, 10 Jun 2026 00:00:00 GMT langchain 1.3.5 adds compound trigger conditions to SummarizationMiddleware so summarization fires only when token count and message count both cross thresholds. https://kmail.at/tutorials/lc-2026-06-09-dict-asdict https://kmail.at/tutorials/lc-2026-06-09-dict-asdict Tue, 09 Jun 2026 00:00:00 GMT LangChain 1.4.2 deprecates .dict() across all objects in favor of .asdict(), fixing a clash with Python\u2019s builtin dict annotation. https://kmail.at/tutorials/lc-2026-06-08-convert-to-message https://kmail.at/tutorials/lc-2026-06-08-convert-to-message Mon, 08 Jun 2026 00:00:00 GMT LangChain core 1.4.1 fixes message round-tripping so Serializable constructor-envelope payloads deserialize back into Message objects.