https://openai.claude-world.com/Thirty-two retrospective technical issues, eighteen Codex CLI lessons, and six OpenAI Platform companion guides.enMon, 10 Aug 2026 00:00:00 GMThttps://openai.claude-world.com/learn/2026-08-chatgpt-app-ui/https://openai.claude-world.com/learn/2026-08-chatgpt-app-ui/The MCP server owns tool truth and authorization; the embedded component renders state and requests actions through a narrow bridge.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveAPPShttps://openai.claude-world.com/learn/2026-07-codex-subagents-ci/https://openai.claude-world.com/learn/2026-07-codex-subagents-ci/Subagents need independent ownership and result contracts; headless CI needs fixed permissions, machine-readable output, and isolated credentials.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveCODEXhttps://openai.claude-world.com/learn/2026-06-codex-mcp-skills/https://openai.claude-world.com/learn/2026-06-codex-mcp-skills/MCP connects authenticated external tools and context; a Skill packages instructions, scripts, references, and a trigger for a reusable workflow.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveCODEXhttps://openai.claude-world.com/learn/2026-05-codex-config-sandbox/https://openai.claude-world.com/learn/2026-05-codex-config-sandbox/Configuration chooses defaults, the sandbox enforces filesystem and network limits, and approvals govern exceptional authority.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveCODEXhttps://openai.claude-world.com/learn/2026-04-agents-md/https://openai.claude-world.com/learn/2026-04-agents-md/Durable guidance should name architecture boundaries, commands, conventions, and verification without replacing the current task.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveCODEXhttps://openai.claude-world.com/learn/2026-03-codex-cli/https://openai.claude-world.com/learn/2026-03-codex-cli/The working directory, task, sandbox, approval policy, and definition of done should be explicit before the coding agent edits files.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveCODEXhttps://openai.claude-world.com/learn/2026-02-realtime-webrtc/https://openai.claude-world.com/learn/2026-02-realtime-webrtc/A server-authenticated session bootstrap lets the browser establish media transport without receiving the long-lived platform key.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveREALTIMEhttps://openai.claude-world.com/learn/2026-01-realtime-architecture/https://openai.claude-world.com/learn/2026-01-realtime-architecture/Transport, session configuration, turn detection, interruption, tools, reconnect, and terminal states jointly define the experience.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveREALTIMEhttps://openai.claude-world.com/learn/2025-12-evals/https://openai.claude-world.com/learn/2025-12-evals/Define representative inputs, human-readable criteria, automated graders where appropriate, and a baseline before changing the system.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveEVALShttps://openai.claude-world.com/learn/2025-11-tracing-observability/https://openai.claude-world.com/learn/2025-11-tracing-observability/Connect run, turn, model, tool, handoff, approval, latency, usage, and terminal outcome with stable correlation identifiers.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveOBSERVEhttps://openai.claude-world.com/learn/2025-10-agent-handoffs/https://openai.claude-world.com/learn/2025-10-agent-handoffs/A handoff carries one goal, bounded context, allowed tools, authority, deadline, output shape, and the next owner.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveAGENTShttps://openai.claude-world.com/learn/2025-09-agents-sdk/https://openai.claude-world.com/learn/2025-09-agents-sdk/The controller must own turns, tools, state, approvals, stop conditions, and traces; specialists receive narrow tasks and typed results.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveAGENTShttps://openai.claude-world.com/learn/2025-08-mcp-connectors/https://openai.claude-world.com/learn/2025-08-mcp-connectors/Inventory server identity, transport, authentication, exposed tools, data returned, approvals, and failure behavior before enabling it.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveMCPhttps://openai.claude-world.com/learn/2025-07-tool-approvals/https://openai.claude-world.com/learn/2025-07-tool-approvals/Classify tools by impact and pause before external messages, purchases, destructive writes, credential use, or privilege changes.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveSAFETYhttps://openai.claude-world.com/learn/2025-06-background-mode/https://openai.claude-world.com/learn/2025-06-background-mode/Background work needs a job identifier, polling or streaming policy, cancellation, expiry, idempotent consumption, and a user-visible status.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveASYNChttps://openai.claude-world.com/learn/2025-05-code-interpreter/https://openai.claude-world.com/learn/2025-05-code-interpreter/Send bounded inputs, define expected artifacts, and validate outputs outside the sandbox before they enter business workflows.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveTOOLShttps://openai.claude-world.com/learn/2025-04-file-search/https://openai.claude-world.com/learn/2025-04-file-search/A hosted retrieval tool still requires corpus governance: ingestion status, metadata, access filters, citations, refresh, and deletion.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveRETRIEVALhttps://openai.claude-world.com/learn/2025-03-web-search/https://openai.claude-world.com/learn/2025-03-web-search/Search results are volatile evidence: preserve source URLs, distinguish retrieval time from event time, and show citations near claims.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveTOOLShttps://openai.claude-world.com/learn/2025-02-conversation-state/https://openai.claude-world.com/learn/2025-02-conversation-state/Decide whether the application resends history, chains response identifiers, or uses a conversation resource, and make that choice visible in data policy.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveSTATEhttps://openai.claude-world.com/learn/2025-01-responses-api/https://openai.claude-world.com/learn/2025-01-responses-api/Typed input and output items let messages, reasoning, tool calls, and tool results remain distinct across a multi-step request.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveRESPONSEShttps://openai.claude-world.com/learn/2024-12-batch-processing/https://openai.claude-world.com/learn/2024-12-batch-processing/Batch work needs stable custom identifiers, immutable input manifests, per-item status, result reconciliation, and bounded replay.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveOPERATIONShttps://openai.claude-world.com/learn/2024-11-embeddings-retrieval/https://openai.claude-world.com/learn/2024-11-embeddings-retrieval/Embeddings support semantic comparison; a production retrieval layer also needs chunk ownership, metadata filters, freshness, citations, and deletion.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveRETRIEVALhttps://openai.claude-world.com/learn/2024-10-text-to-speech/https://openai.claude-world.com/learn/2024-10-text-to-speech/Choose voice, format, chunking, playback, and interruption behavior as product decisions, then disclose that the voice is AI-generated.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveAUDIOhttps://openai.claude-world.com/learn/2024-09-speech-to-text/https://openai.claude-world.com/learn/2024-09-speech-to-text/Treat transcription as a media-processing job with upload validation, language context, segmentation, speaker policy, and correction workflow.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveAUDIOhttps://openai.claude-world.com/learn/2024-08-image-generation/https://openai.claude-world.com/learn/2024-08-image-generation/Separate prompt intent, generation request, moderation state, asset storage, metadata, and editorial acceptance into observable stages.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveMEDIAhttps://openai.claude-world.com/learn/2024-07-vision-input/https://openai.claude-world.com/learn/2024-07-vision-input/Define what the model may infer from an image, preserve the original asset reference, and separate observation from downstream decisions.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveVISIONhttps://openai.claude-world.com/learn/2024-06-streaming-ui/https://openai.claude-world.com/learn/2024-06-streaming-ui/A useful stream distinguishes start, typed deltas, tool activity, completion, cancellation, and failure rather than appending arbitrary text.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveUXhttps://openai.claude-world.com/learn/2024-05-function-calling/https://openai.claude-world.com/learn/2024-05-function-calling/A tool definition expresses callable intent; authenticated server code must own validation, execution, idempotency, and the returned result.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveTOOLShttps://openai.claude-world.com/learn/2024-04-structured-outputs/https://openai.claude-world.com/learn/2024-04-structured-outputs/Structured Outputs reduce parsing ambiguity, but product code must still handle refusals, incomplete results, and domain-invalid values.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveDATAhttps://openai.claude-world.com/learn/2024-03-model-routing/https://openai.claude-world.com/learn/2024-03-model-routing/Describe each workload by modality, tool support, latency, quality, and budget, then resolve that profile to a deployable model in configuration.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveMODELShttps://openai.claude-world.com/learn/2024-02-prompt-contracts/https://openai.claude-world.com/learn/2024-02-prompt-contracts/Separate durable developer instructions from user input, give the task an explicit output contract, and keep prompt changes reviewable.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archivePROMPTShttps://openai.claude-world.com/learn/2024-01-api-request-boundary/https://openai.claude-world.com/learn/2024-01-api-request-boundary/Start with one server-owned adapter for authentication, request IDs, timeouts, and normalized errors instead of calling the platform throughout the UI.Mon, 10 Aug 2026 00:00:00 GMTMonthly technical archiveAPIhttps://openai.claude-world.com/guides/install-and-diagnose/https://openai.claude-world.com/guides/install-and-diagnose/A reliable Codex workflow starts before the first prompt. Install from the official standalone script, identify the binary you are actually running, update deliberately, and use the built-in diagnostic report instead of guessing about configuration, authentication, Git, or terminal failures.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/authentication-and-credentials/https://openai.claude-world.com/guides/authentication-and-credentials/Codex supports ChatGPT sign-in for subscription access and API-key sign-in for usage-based local work. The safe path is to choose the flow deliberately, pipe secrets through stdin, verify the active method, and treat local auth storage like a password.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/first-terminal-session/https://openai.claude-world.com/guides/first-terminal-session/The first useful Codex session begins in the right repository with a bounded permission mode and a prompt that states goal, context, constraints, and done condition. Codex can explore, edit, run checks, and report evidence, but only when the task boundary is explicit.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/sandbox-approvals-network/https://openai.claude-world.com/guides/sandbox-approvals-network/Codex security is the combination of an OS-enforced sandbox and an approval policy. Workspace-write with on-request approvals is a practical local default; read-only is for inspection and planning. Web search does not automatically grant network access to spawned commands.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/interactive-session-control/https://openai.claude-world.com/guides/interactive-session-control/Codex CLI exposes keyboard-first controls for live work plus top-level commands for saved-session lifecycle. Plan mode improves ambiguous work; status and diff expose state; compact preserves key context; review creates a dedicated inspection pass. Resume, fork, archive, unarchive, and permanent delete keep persisted work explicit.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/agents-md-guidance/https://openai.claude-world.com/guides/agents-md-guidance/Codex reads an instruction chain before work begins: global guidance, then project files from repository root toward the working directory. Closer guidance appears later and overrides broader guidance. AGENTS.override.md wins at a level, and the combined project guidance has a default size limit.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/config-profiles-models/https://openai.claude-world.com/guides/config-profiles-models/Model-aware planning is more than selecting the largest model. Keep the CLI that already holds the repository and conversation context, map each task to a role, select only an effort advertised by that model, and keep subagent orchestration separate from single-task depth. Codex configuration still resolves by scope, while model access can drift by release, account, authentication method, quota, and rollout.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/diff-test-review/https://openai.claude-world.com/guides/diff-test-review/Codex can implement, run checks, show the working-tree diff, and launch a dedicated reviewer. Review scopes are mutually exclusive: uncommitted work, a base-branch comparison, a commit, or custom instructions. The reviewer reports findings without changing the tree.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/mcp-connections/https://openai.claude-world.com/guides/mcp-connections/Codex supports local STDIO and Streamable HTTP MCP servers. CLI, IDE extension, and desktop app share host configuration. A reliable connection names its transport, authentication source, timeouts, required status, enabled tools, and approval behavior before a workflow depends on it.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/skills-and-plugins/https://openai.claude-world.com/guides/skills-and-plugins/Skills package instructions, references, scripts, and optional assets. Codex discovers them through concise metadata and loads full instructions only when selected. Plugins distribute skills and optional MCP connections through configured marketplaces; list and review a marketplace before installing a named plugin.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/hooks-and-rules/https://openai.claude-world.com/guides/hooks-and-rules/Hooks run trusted command handlers at lifecycle events such as PreToolUse, PostToolUse, Stop, and SessionStart. Rules decide whether matching command prefixes are allowed, prompted, or forbidden outside the sandbox. Hooks are stable; rules remain experimental and the most restrictive matching decision wins.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/subagents-exec-ci/https://openai.claude-world.com/guides/subagents-exec-ci/Subagents protect the main thread from noisy exploration and can run independent read-heavy work in parallel. For non-interactive workflows, codex exec defaults to a read-only sandbox, streams progress to stderr, and prints the final message to stdout; JSONL and output schemas support machine consumers. GitHub Actions should use the official action and narrow privileges.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/exec-headless-automation/https://openai.claude-world.com/guides/exec-headless-automation/codex exec is the non-interactive entry point for scripts and CI. It accepts a prompt as an argument or stdin, uses the same repository guidance and sandbox model as interactive Codex, returns a meaningful process status, and can resume saved exec sessions. Reliability comes from a bounded working directory, explicit permissions, isolated credentials, and verification that does not depend on a conversational transcript.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/structured-output-jsonl/https://openai.claude-world.com/guides/structured-output-jsonl/With --json, codex exec emits newline-delimited JSON events to stdout. --output-schema constrains the final response with a JSON Schema, while --output-last-message writes the final message to a chosen file. These are distinct surfaces: event streaming supports progress and diagnostics; the final artifact supports a stable consumer contract. Neither proves the underlying code change is correct.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/session-lifecycle/https://openai.claude-world.com/guides/session-lifecycle/Codex persists interactive and exec sessions so work can continue later. resume continues a lineage; fork creates a branch; archive and unarchive provide reversible housekeeping; delete permanently removes the selected saved session. The safe unit of continuation is the pair of an exact session and a compatible repository state.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/cloud-tasks-and-apply/https://openai.claude-world.com/guides/cloud-tasks-and-apply/The experimental cloud command can submit a task to a named environment, list or inspect task state, display a selected attempt’s unified diff, and apply that diff to the local working tree. Applying is not merging and does not prove the patch fits the current branch. Inspect the exact task, attempt, local status, and resulting tests before accepting anything.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/sandbox-command-and-execpolicy/https://openai.claude-world.com/guides/sandbox-command-and-execpolicy/codex sandbox runs a command inside a Codex-provided sandbox and can resolve a named permission profile, narrow readable roots or network, and log macOS denials. execpolicy check evaluates a command against rule files and explains the decision. These tools diagnose and test authority; they do not make an unsafe command safe or replace operating-system enforcement.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/diagnostics-features-and-completions/https://openai.claude-world.com/guides/diagnostics-features-and-completions/Codex provides doctor for layered installation and runtime health, features for staged feature state, completion for shell integration, and debug utilities for targeted evidence such as the bundled model catalog. Start with read-only inspection, preserve version context, and review generated or diagnostic output before placing it in tickets or startup files.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/responses-boundary/https://openai.claude-world.com/guides/responses-boundary/The Responses API is the recommended starting point for new OpenAI integrations. Treat it as an application boundary—not a thin text-generation wrapper—so messages, tool calls, tool results, and future input types remain explicit.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/tool-contracts/https://openai.claude-world.com/guides/tool-contracts/OpenAI tools can search, retrieve files, execute code, connect through MCP, or call your own functions. Reliability comes from the boundary around each action: a small schema, explicit authority, idempotency, and a result the model can interpret.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/structured-outputs/https://openai.claude-world.com/guides/structured-outputs/Structured Outputs can constrain model responses to a supplied JSON Schema. That removes a class of parsing failures, but it does not prove that values are true, useful, authorized, or safe to apply.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/agent-loops/https://openai.claude-world.com/guides/agent-loops/An agent plans, calls tools, carries state, and may hand work to specialists. The key architecture decision is not how many agents exist—it is who owns the run loop, tool policy, state, approvals, and trace.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/realtime-systems/https://openai.claude-world.com/guides/realtime-systems/Realtime sessions are designed for low-latency live interactions. A production voice experience is more than streaming audio: it needs a transport chosen for the client, explicit session state, interruption behavior, tool latency budgets, and recovery from partial turns.Fri, 31 Jul 2026 00:00:00 GMThttps://openai.claude-world.com/guides/model-portability/https://openai.claude-world.com/guides/model-portability/The OpenAI model catalog changes over time. Product code should ask for a tested capability profile—quality, latency, modalities, tool support, and budget—then resolve that profile through configuration and evaluations.Fri, 31 Jul 2026 00:00:00 GMT