# Experiential Labs: machine-readable reference Experiential Labs is an OpenAI-compatible model gateway: one base URL in front of every model, spanning hosted providers, your own provider keys (bring-your-own-key), our platform-funded credits, and self-hosted or custom models, plus a public model catalog, per-model provider waterfalls, organization API keys, and usage and credits. Point any OpenAI client at this gateway and change nothing else. This file is the complete agent-facing reference. Human docs: https://pr-683.preview.experientiallabs.ai/docs. OpenAI-shared semantics: https://platform.openai.com/docs/api-reference ## Base URLs (local vs platform) This deployment: API base URL: https://api-pr-683.preview.experientiallabs.ai Web app: https://pr-683.preview.experientiallabs.ai This is a self-hosted or local deployment. The hosted platform is https://api.experientiallabs.ai (web: https://platform.experientiallabs.ai). OpenAI clients use base_url = "https://api-pr-683.preview.experientiallabs.ai/v1". The management API is under https://api-pr-683.preview.experientiallabs.ai/api. ## Authentication - One header, always: Authorization: Bearer . No X-Api-Key, no query param. - Keys look like xpl_ + 40 lowercase hex chars and are scoped to ONE organization. The secret is shown once, at creation. - Mint a key by signing in to https://pr-683.preview.experientiallabs.ai/settings/api-keys; the plaintext appears exactly once. Key creation and revocation are web-session actions, not API-key actions. - One key reaches BOTH the inference surface (/v1/*) and the management API an agent needs: catalog reads, custom-model and waterfall writes, BYOK provider connections, usage reads, and the org's key list. It CANNOT mint or revoke keys, change another key's limits, or reach platform-admin routes. - Bad or missing key: 401 with {"error":{"code":"invalid_key", "type":"authentication_error"}}. The 401 does not distinguish absent, malformed, revoked, or expired. - Verify a key works: GET https://api-pr-683.preview.experientiallabs.ai/v1/models with the key -> 200 and the org's callable models. ## Two lanes (how a call is paid for) - pass_through: your own provider key (BYOK). No markup: the provider bills you directly. Connect keys at https://pr-683.preview.experientiallabs.ai/settings. - platform_funded: our credits. Public-catalog models are priced from the launch catalog; each call draws down your credit balance. No markup. Which lane a model uses is decided by its provider waterfall. Either way, the gateway adds zero markup. ## Models - GET https://api-pr-683.preview.experientiallabs.ai/v1/models lists the model slugs your key can call: the public catalog plus your organization's own custom and local models. Each id is a slug, e.g. "claude-opus-5", "gpt-5.5", "gemini-3.7-flash". - A slug resolves through a provider waterfall: an ordered list of ways to reach the model (provider + provider model id). The gateway tries each rung in order, fails over on capacity and transport errors, and returns the first success. Organizations may override the default chain. - The full catalog with pricing, context window, and modalities is a public, keyless read: GET https://api-pr-683.preview.experientiallabs.ai/api/models (no key returns the public rows; send your key to also see your org's own custom and local models). One model: GET https://api-pr-683.preview.experientiallabs.ai/api/models/; its deployments: .../providers. Note the split: the /api/models* catalog reads are keyless, while GET https://api-pr-683.preview.experientiallabs.ai/v1/models (the OpenAI-compatible list) requires your key. ## OpenAI-compatible inference API GET https://api-pr-683.preview.experientiallabs.ai/v1/models -> {"object": "list", "data": [{"id": "", "object": "model", "owned_by": "exp", ...}]} Scoped to the key's organization. POST https://api-pr-683.preview.experientiallabs.ai/v1/chat/completions Standard OpenAI Chat Completions. "model" MUST be a slug from /v1/models. "stream": true streams as SSE. POST https://api-pr-683.preview.experientiallabs.ai/v1/responses Standard OpenAI Responses. "stream": true streams as SSE. "previous_response_id" continues a prior response on any worker instance; continuations are retained for 24 hours. An unknown or expired id -> 400 code=continuation_unavailable: resend the full conversation. POST https://api-pr-683.preview.experientiallabs.ai/v1/messages Anthropic Messages API, translated onto the same chat surface and models. Auth: x-api-key OR Authorization: Bearer (same xpl_ key). "stream": true streams Anthropic SSE events. Limits of the translation lane: extended thinking is unavailable (thinking config and blocks are accepted and dropped), image/document blocks -> 400 (the chat surface is text-only), Idempotency-Key is not honored here, /v1/messages/count_tokens answers an explicit 404 not_found_error (estimate locally), and errors use Anthropic's envelope {"type":"error","error":{"type","message"}} at the same statuses as below. - Idempotency-Key is honored on the OpenAI routes: an exact retry with the same key replays the original result; the same key with a different body -> 409 idempotency_conflict. - Every other /v1/* path answers 404 code=not_found in the OpenAI error envelope. ## Error envelope (all /v1 routes; /v1/messages wraps the same statuses and ## meanings in Anthropic's envelope instead) Errors are {"error": {"message", "type", "code", "param"?}}. Stable codes (code -> HTTP status -> meaning -> recovery): invalid_json 400 body is not valid JSON -> fix the request. invalid_request 400 malformed request -> read message, fix, resend. invalid_parameter 400 a field is invalid ("param" names it) -> fix it. unsupported_capability 400 the model cannot do what you asked (a tool, a modality, reasoning) -> pick a capable model; check supported_params and modalities in /api/models. continuation_unavailable 400 previous_response_id is unknown or expired -> resend the full conversation. invalid_key 401 missing/bad/expired/revoked key -> fix Authorization. model_not_granted 403 your org cannot call this slug -> use one from /v1/models. idempotency_conflict 409 same Idempotency-Key, different body -> new key. idempotency_replay_unavailable 409/500 original keyed result gone after a restart -> resend with a new Idempotency-Key. insufficient_quota 429 a spend limit or your credit balance is exhausted (message says which: a daily org or per-model cap, or credits) -> add credits or raise limits at https://pr-683.preview.experientiallabs.ai/credits (platform-funded lane). unavailable_route 429/503 throttled or no healthy route right now -> retry with backoff. gateway_overloaded 429 -> retry with backoff. request_cancelled 499 the client disconnected before completion. all_routes_failed 502 every provider in the waterfall failed -> retry; if BYOK, check your provider key. provider_output_too_large 502 -> lower max output tokens. gateway_draining 503 instance is draining -> retry (hits another). deadline_exceeded 504 request ran past the deadline -> shorten or retry. internal_error 500 -> retry with backoff. Retry 429 (throttled)/502/503/504 with backoff. Do NOT blindly retry 400/401/403/409: fix the request first. Retries can double-bill a provider (at-least-once); pass an Idempotency-Key to dedupe. ## Management API (same Bearer key; https://api-pr-683.preview.experientiallabs.ai/api) Catalog reads (keyless-public: public rows without a key, plus your org's own rows when you send your key): GET /api/models[?modality=&category=&provider=&min_context=&sort=&limit=&offset=] GET /api/models/ GET /api/models//providers Custom and local models and waterfalls (the key acts for its own org): POST /api/models create a custom model: {slug, display_name, providers:[{provider, provider_model_id, base_url?, ...}]} POST /api/models//providers add a deployment or local variant GET/PUT /api/models//waterfall read or replace the ordered chain ({model_provider_ids:[...]}) BYOK provider connections: GET /api/orgs//provider-connections PUT /api/orgs//provider-connections/ connect/rotate ({secret, config}) POST /api/orgs//provider-connections//check verify Usage and keys: GET /api/gateway/usage/daily?org_id=&scope=&group_by= rollup (day|model|member) GET /api/gateway/usage/events?org_id=... per-request stream GET /api/gateway/catalog?org_id=... aliases as your org resolves them, with each one's lane GET /api/keys the org's keys (never secrets) Archived Projects: The trace-backed optimizer, router-building, and Project-serving surfaces are retained for historical data and operator recovery only. Their UI routes and customer-key build paths are closed; do not use them as an inference API or as an alternate modeling implementation. Trace telemetry (bring your traces in as telemetry only; never a router build; the key acts for its org): POST /api/orgs//telemetry/traces/pull live pull from a provider: {transport_kind, source_kind, source_label, credential, config?}. transport_kind is one of: braintrust, langsmith, langfuse, posthog, mastra, postgres. The credential is used once and never stored on the row. POST /api/orgs//telemetry/traces/upload multipart {source_kind, source_label, file}; same formats and <=50MB limit as the Project upload. Arize/Phoenix have no live pull yet — use this with source_kind phoenix (or otlp). GET /api/orgs//telemetry/traces the org's landed telemetry traces plus total_ingests and total_traces (the verify-count). Humans see these at https://pr-683.preview.experientiallabs.ai/telemetry. This path never creates a Project, preparation, or optimize job. Providers accepted on a deployment: openai, anthropic, gemini, azure_openai, openrouter, bedrock, local, fireworks, modal, experiential_cloud. Experiential Cloud is a curated collection of models, hosted and optimized by Experiential Labs. Call it with an xpl_ key. Organizations do not connect their own credentials to it. BYOK connections stay: openai, anthropic, gemini, azure_openai, openrouter, bedrock, fireworks, modal. ## Coding agents (Claude Code, Conductor, Codex, OpenCode, Cline, and any ## OpenAI-compatible tool) Any coding agent that can target an OpenAI-compatible endpoint works against this gateway: set its base URL to https://api-pr-683.preview.experientiallabs.ai/v1, supply an xpl_ key as the Bearer token, and name models by slug (from GET https://api-pr-683.preview.experientiallabs.ai/v1/models). Tools that read the standard OpenAI SDK environment variables need only: export OPENAI_BASE_URL="https://api-pr-683.preview.experientiallabs.ai/v1" export OPENAI_API_KEY="xpl_..." Claude Code connects through the Anthropic Messages lane: export ANTHROPIC_BASE_URL="https://api-pr-683.preview.experientiallabs.ai" export ANTHROPIC_AUTH_TOKEN="xpl_..." export ANTHROPIC_MODEL="" Any catalog slug works as the model, not just Claude models. See the /v1/messages limits above (no extended thinking, no images). Conductor (parallel Claude Code agents) uses the same variables in its Settings -> Environment (Claude Code section), or per-repo in .conductor/settings.local.toml under [environment_variables]; additionally set ANTHROPIC_API_KEY to the EMPTY string there so Claude Code never tries to authenticate with Anthropic directly. Verified per-agent configs (Claude Code, OpenAI Codex CLI over the Responses API, OpenCode, Cline) are maintained at https://pr-683.preview.experientiallabs.ai/docs/coding-agents. ## Self-hosted CLI Self-hosters run the open-source Experiential gateway from the terminal with "exp run". The hosted platform manages the catalog, keys, and usage for you in the web app. ## Typical agent flow 1. A human signs in and mints an API key at https://pr-683.preview.experientiallabs.ai/settings/api-keys, and (for the pass-through lane) connects provider keys at https://pr-683.preview.experientiallabs.ai/settings. 2. The agent receives the key. It calls GET https://api-pr-683.preview.experientiallabs.ai/v1/models to see callable slugs. 3. The agent calls POST https://api-pr-683.preview.experientiallabs.ai/v1/chat/completions (or /v1/responses) with model="" exactly as it would call OpenAI, streaming or not. 4. The agent reads its own usage and spend at GET https://api-pr-683.preview.experientiallabs.ai/api/gateway/usage/daily; humans see the same at https://pr-683.preview.experientiallabs.ai/telemetry and https://pr-683.preview.experientiallabs.ai/credits. ## Setup prompts (paste one into a coding agent; it does the setup for you) These are the same first-person prompts the web app's onboarding uses, with this deployment's URLs already filled in. A human pastes one into a CLI coding agent; the agent follows it literally. Human-readable copies: https://pr-683.preview.experientiallabs.ai/docs/setup-prompts. Shareable source copies: https://github.com/experientiallabs/setup-prompts. ### Create an account from your coding agent Paste into a CLI agent to create your account instantly from your email, wire the gateway, and confirm the key. 1. Ask me for my email address — it is the one thing signup needs. Ask me directly ("What's your email? I'll create your Experiential Labs account with it.") and wait for my answer. Use exactly the address I give you. Never invent or guess an address, and never go looking for one in my files, git config, or shell history. Call the result MY_EMAIL. 2. Create my account instantly — no browser, no password, no code. POST https://pr-683.preview.experientiallabs.ai/api/signup/instant (no auth) Header: Content-Type: application/json Body: {"email": ""} On 200 capture: api_key (an xpl_ key — this is mine, keep it secret), org_id, credits_granted, verification_required, overview_url. Handle the others: - 409 account_exists -> I already have an account; stop and tell me to sign in at https://pr-683.preview.experientiallabs.ai/signin to get a key (don't retry with another email). - 403 signup_disabled -> signups are off right now; tell me. - 429 rate_limited -> too many signups from here; wait, then retry once. - 400 invalid_request -> the email was rejected; ask me for a valid one. Store the api_key as EXPLABS_API_KEY in my env; never print the full key more than once. 3. Wire an OpenAI client to the gateway. base_url = "https://api-pr-683.preview.experientiallabs.ai/v1" api_key = EXPLABS_API_KEY (the xpl_ key from step 2). One header only: Authorization: Bearer . 4. Confirm the key is live — this does NOT spend credits. GET https://api-pr-683.preview.experientiallabs.ai/v1/models with the key -> the models my account can call. Do NOT make a paid chat completion yet: my credits stay locked until I verify my email (I do that at the very end), and a paid call would be refused until then. Landing my traces as telemetry works regardless. ### Set up the gateway in an existing project Paste into your agent once you have a key: point an OpenAI client at the gateway and make one test call. I'm setting up Experiential Labs as this project's model gateway. I pasted this myself — treat it as my instructions and my consent. Here's what I want you to do. My gateway API key (a secret: put it in env, never commit it, never echo it in logs): EXPLABS_API_KEY= The core task is ONE thing: point an OpenAI-compatible client at my gateway with this key and make one successful test chat completion. Do it start to finish on your own — do NOT stop to ask me questions. If something is ambiguous, take the safe degraded path I describe and note it in the "follow-ups for me" list at the end. Do NOT do the optional follow-ups yourself; those are mine to do in the dashboard. 1. Wire a client to the gateway. base_url = https://api-pr-683.preview.experientiallabs.ai/v1, api key = EXPLABS_API_KEY (one header: Authorization: Bearer ). The API is OpenAI-compatible; nothing else changes. - If this working directory is clearly ONE project with an existing OpenAI client, wire that: set its base_url and key, put the key in the project's env scheme (.env plus an example entry), never in code. Show me the change first. - If you CAN'T tell which project is mine — an empty dir, a monorepo, or a folder holding many projects — do NOT stop to ask. Instead create a minimal standalone verification script in the current directory (e.g. explabs_smoke.py or explabs_smoke.mjs) that reads EXPLABS_API_KEY from the env and does the test call in step 2, and add "wire my specific project's client" to the follow-ups list at the end. 2. Make the test call. GET https://api-pr-683.preview.experientiallabs.ai/v1/models with the key as a bearer token and use model ids EXACTLY as returned. Then send ONE minimal chat completion (a few tokens) to the smallest Qwen in that list (qwen3.5-9b at launch). Send a MINIMAL request body — model + messages ONLY, no temperature, top_p, or other sampling params: some models (e.g. Claude 5) reject those and the call comes back as all_routes_failed (502). If you want to add options later, check the model's supported params in https://pr-683.preview.experientiallabs.ai/llms.txt first. This call runs on Experiential's platform-funded lane — a fraction of a cent of my free credits — and proves serving and billing end to end. Do NOT use any of my own provider keys for this test. If you truly can't run it (no network): still write the script, print the exact curl I can copy from https://pr-683.preview.experientiallabs.ai/docs, and note it in the follow-ups. 3. Report back, then hand me the follow-ups. Tell me: what you changed (or the standalone script you created), and the test call's model, that it succeeded, and its cost. Then list "follow-ups for me" — anything you deferred above, PLUS these optional steps, which are MINE to do in the dashboard (do NOT attempt them yourself, and don't block on them): - Wire my specific project's client, if you only created a standalone script (skip this line if you already wired a real project). - Connect my own provider keys (BYOK) for free pass-through. I'll add them at https://pr-683.preview.experientiallabs.ai/settings/integrations — verified on save, confirmed per key by prefix, landing in my own gateway account and nowhere else. - Import my existing Claude Code / Codex usage history (metadata only) so my dashboard opens with real spend, from https://pr-683.preview.experientiallabs.ai/settings. - Review all my balances and spend — platform-funded credits and free pass-through both — at https://pr-683.preview.experientiallabs.ai/credits, summarized on my Overview. Read the contract if you build further: https://pr-683.preview.experientiallabs.ai/docs (human docs) and https://pr-683.preview.experientiallabs.ai/llms.txt (machine-readable: honored and refused parameters, error codes, streaming caveats). Follow it literally. ### Wire your coding agent to the gateway Paste into Claude Code, Codex, OpenCode, Cline, Conductor, or any OpenAI-SDK tool: the agent identifies itself and applies its own verified integration. I pasted this into you myself — wire THIS coding agent up to my Experiential Labs gateway, so my model calls route through https://api-pr-683.preview.experientiallabs.ai and show up in my usage. Ground rules: never print my full key (first 8 characters at most), ask me before you edit any config file or shell profile, and if you cannot do a step (no file access, settings live in a GUI), print the exact manual steps for me instead. Print what you're doing at each step. 1. Get my key. Use EXPLABS_API_KEY from my environment if it is set; otherwise ask me to paste one (I can mint it at https://pr-683.preview.experientiallabs.ai/settings/api-keys). It looks like xpl_ followed by 40 hex characters. 2. Prove the key works before touching any config: GET https://api-pr-683.preview.experientiallabs.ai/v1/models with header "Authorization: Bearer " -> 200 and the model slugs I can call. Remember the list; I'll pick models from it. 3. Identify which agent you are and apply YOUR integration: - Claude Code: You cannot repoint yourself mid-session. Offer me two placements and apply the one I pick: (a) append to my shell profile, or (b) print for one-off use: export ANTHROPIC_BASE_URL="https://api-pr-683.preview.experientiallabs.ai" # no /v1 suffix export ANTHROPIC_AUTH_TOKEN="" export ANTHROPIC_MODEL="" Then tell me to restart you from a shell with those set. Warn me plainly: extended thinking is unavailable through the gateway and image pastes are rejected (the lane is text-only). - Conductor: Print these for me to put in Settings -> Environment (Claude Code section), or write them to this repo's .conductor/settings.local.toml under [environment_variables] if I prefer (make sure that file is git-ignored before writing my key into it): ANTHROPIC_BASE_URL = "https://api-pr-683.preview.experientiallabs.ai" ANTHROPIC_AUTH_TOKEN = "" ANTHROPIC_API_KEY = "" The empty ANTHROPIC_API_KEY is required: it stops Claude Code from trying to authenticate with Anthropic directly. - OpenAI Codex CLI: Add to ~/.codex/config.toml (create it if missing, show me the diff before writing): model = "" model_provider = "explabs" [model_providers.explabs] name = "Experiential Labs" base_url = "https://api-pr-683.preview.experientiallabs.ai/v1" env_key = "EXPLABS_API_KEY" wire_api = "responses" Leave requires_openai_auth unset, and make sure EXPLABS_API_KEY is exported where I launch you. - OpenCode: Write the provider block into this project's opencode.json (or ~/.config/opencode/opencode.json if I prefer global — ask): {"provider": {"explabs": {"npm": "@ai-sdk/openai-compatible", "name": "Experiential Labs", "options": {"baseURL": "https://api-pr-683.preview.experientiallabs.ai/v1", "apiKey": "{env:EXPLABS_API_KEY}"}, "models": {"": {"name": ""}}}} Fill limit.context/limit.output for each slug from GET https://api-pr-683.preview.experientiallabs.ai/api/models/ so my context window is right. - Cline: Your settings live in the VS Code UI, so print these for me to set by hand: API Provider "OpenAI Compatible"; Base URL https://api-pr-683.preview.experientiallabs.ai/v1; API Key = my key (no Bearer prefix); Model ID = a slug from step 2; and per-model context window / max output tokens from https://api-pr-683.preview.experientiallabs.ai/api/models/. - Any other OpenAI-compatible tool: Export OPENAI_BASE_URL="https://api-pr-683.preview.experientiallabs.ai/v1" and OPENAI_API_KEY="" wherever I launch you, and name models by slug. If your own config wants the values instead, it needs the same three: base URL https://api-pr-683.preview.experientiallabs.ai/v1, my key, and a slug. 4. Verify end to end. Make one tiny completion with the surface you configured ("reply with the single word: ok", small max output) — via your own next model call if you now route through the gateway, otherwise via curl with my key. Then tell me it landed and that I can watch every call at https://pr-683.preview.experientiallabs.ai/telemetry. 5. Recap exactly what you changed (files and values, key shown as xpl_ prefix only) so I can undo it later. ### Bring your traces in as telemetry Paste into a CLI agent to create your account instantly and land your existing LLM traces as telemetry. I pasted this into you myself — please create an Experiential Labs account for me and then pull my existing LLM traces onto the platform as telemetry. This is my consent to create the account under my identity (using the email I give you when you ask), to read my own trace files, and to use the observability credential I give you. No browser step is needed to start — my account is created instantly from my email; I only verify that email later to unlock my credits, and I don't need to for my traces to land. This uploads my existing traces so I can see them as telemetry on the platform. Do it in this order. Print what you're doing at each step. If you lack a capability (no network, no file access) or get stuck, stop and tell me exactly what to do manually. Never invent an email, an API key, a credential, or a file path — if you don't have one, ask me. 1. Ask me for my email address — it is the one thing signup needs. Ask me directly ("What's your email? I'll create your Experiential Labs account with it.") and wait for my answer. Use exactly the address I give you. Never invent or guess an address, and never go looking for one in my files, git config, or shell history. Call the result MY_EMAIL. 2. Create my account instantly — no browser, no password, no code. POST https://pr-683.preview.experientiallabs.ai/api/signup/instant (no auth) Header: Content-Type: application/json Body: {"email": ""} On 200 capture: api_key (an xpl_ key — this is mine, keep it secret), org_id, credits_granted, verification_required, overview_url. Handle the others: - 409 account_exists -> I already have an account; stop and tell me to sign in at https://pr-683.preview.experientiallabs.ai/signin to get a key (don't retry with another email). - 403 signup_disabled -> signups are off right now; tell me. - 429 rate_limited -> too many signups from here; wait, then retry once. - 400 invalid_request -> the email was rejected; ask me for a valid one. Store the api_key as EXPLABS_API_KEY in my env; never print the full key more than once. 3. Wire an OpenAI client to the gateway. base_url = "https://api-pr-683.preview.experientiallabs.ai/v1" api_key = EXPLABS_API_KEY (the xpl_ key from step 2). One header only: Authorization: Bearer . 4. Confirm the key is live — this does NOT spend credits. GET https://api-pr-683.preview.experientiallabs.ai/v1/models with the key -> the models my account can call. Do NOT make a paid chat completion yet: my credits stay locked until I verify my email (I do that at the very end), and a paid call would be refused until then. Landing my traces as telemetry works regardless. Now bring my traces in as telemetry. Use EXPLABS_API_KEY (the xpl_ key from step 2) as the bearer token on every call below. org_id is from step 2. 5. Interview me: where do my LLM traces live? Ask me one clear question and wait for my answer. Map my answer to exactly one path: - A supported observability provider or database Platform can pull directly (transport_kind one of: braintrust, langsmith, langfuse, posthog, mastra, postgres) -> go to step 6 (live pull). - An exported trace FILE on disk — a raw OpenTelemetry/OTLP export, an Arize/Phoenix export, or any of these upload formats (source_kind one of: otlp, otel-genai, phoenix, braintrust, langfuse, langsmith, mastra, posthog, chat-json) -> go to step 7 (file upload). If I'm on Arize or Phoenix, there's no live pull yet: ask me to export my traces to a file and take the upload path with source_kind "phoenix" (or "otlp" for a raw OpenTelemetry export). If I name a provider that isn't in either list, tell me and offer the file-upload path. Only follow the ONE path that matches my answer. 6. Live pull path — connect the provider and pull. Ask me for the credential for the provider I named (for Braintrust an API key; for LangSmith/Langfuse their API key; for a Postgres database a DSN), plus the small bits of config that provider needs (e.g. Braintrust: the project name; LangSmith/Langfuse: optionally the project/host; Postgres: the table). Then: POST https://api-pr-683.preview.experientiallabs.ai/api/orgs//telemetry/traces/pull Header: Authorization: Bearer $EXPLABS_API_KEY, Content-Type: application/json Body: {"transport_kind": "", "source_kind": "", "source_label": "", "credential": "", "config": {"project": ""}} The credential is used once to pull and is not echoed back. On 201 the response is {"ingest_id", "trace_count", "byte_size", "sha256", ...} — capture trace_count and show it to me. A 400 means bad credentials or config (tell me exactly what it said); 429 means the provider rate-limited us (wait and retry). Then go to step 8. Example (Braintrust): curl -sS -X POST https://api-pr-683.preview.experientiallabs.ai/api/orgs//telemetry/traces/pull \ -H "Authorization: Bearer $EXPLABS_API_KEY" \ -H "Content-Type: application/json" \ -d '{"transport_kind":"braintrust","source_kind":"braintrust", "source_label":"braintrust-prod","credential":"", "config":{"project":""}}' 7. File upload path — find my trace export and upload it. I'm authorizing you to look for my own trace exports: search the current project, ./traces, ./logs, ./data, and my common cache/config paths for JSON or JSONL files that hold LLM/agent spans or runs (names like traces.jsonl, *.otel.jsonl, spans.json, otlp*.json). Show me the candidate files (path + size) and which format each looks like before you upload anything, and map each to ONE source_kind from: otlp, otel-genai, phoenix, braintrust, langfuse, langsmith, mastra, posthog, chat-json (raw OpenTelemetry GenAI spans -> otel-genai; a raw OTLP export -> otlp; an Arize/Phoenix export -> phoenix; a vendor export -> its own name; a plain chat transcript -> chat-json). The file must be UTF-8 JSON or JSONL and at most 50 MB. Then: POST https://api-pr-683.preview.experientiallabs.ai/api/orgs//telemetry/traces/upload This is a multipart/form-data upload with three fields: source_kind = the format you chose source_label = a short label (NOT a file path), e.g. "prod-otel-august" file = the trace file itself On 201 the response is {"ingest_id", "trace_count", "byte_size", "sha256"} — capture trace_count and show it to me. A 422 means the bytes or label failed validation (not JSON/JSONL, empty, or a path-like label) — tell me what it said. Then go to step 8. Example: curl -sS -X POST https://api-pr-683.preview.experientiallabs.ai/api/orgs//telemetry/traces/upload \ -H "Authorization: Bearer $EXPLABS_API_KEY" \ -F source_kind=otlp \ -F source_label=prod-otel-august \ -F file=@ 8. Verify the traces landed as telemetry. GET https://api-pr-683.preview.experientiallabs.ai/api/orgs//telemetry/traces Header: Authorization: Bearer $EXPLABS_API_KEY -> 200 with {"traces": [...], "total_ingests", "total_traces"}. Confirm total_traces is greater than zero and matches the trace_count you captured, and that my ingest is in the list. Tell me the number. When you're done, report back to me: my org_id, whether you pulled live or uploaded a file (and from which provider/format), the ingest_id and its trace_count, and the total_traces the verify read returned. My traces are now telemetry on the platform. I can see them in my dashboard at https://pr-683.preview.experientiallabs.ai/telemetry, and the machine-readable contract is https://pr-683.preview.experientiallabs.ai/llms.txt. One last thing to tell me, then you're done: check my email inbox for a message from Experiential Labs and click the verification link (or enter the code at https://pr-683.preview.experientiallabs.ai/signin) to confirm my email. That's the ONLY thing left for me to do, and it just unlocks my credits for paid model calls — my traces are already live and don't need it. Until I verify, credit-drawing model calls are refused with "insufficient_quota"; using my own provider keys (BYOK) is unaffected. ## Web app URL map (where to send a human) https://pr-683.preview.experientiallabs.ai/ sign in, then land on your key and credits https://pr-683.preview.experientiallabs.ai/models the catalog: search, detail, compare https://pr-683.preview.experientiallabs.ai/playground chat with any model in the browser https://pr-683.preview.experientiallabs.ai/settings/api-keys mint and revoke API keys https://pr-683.preview.experientiallabs.ai/settings connect BYOK provider keys https://pr-683.preview.experientiallabs.ai/credits balance, spend, and adding credits https://pr-683.preview.experientiallabs.ai/telemetry usage by model and by agent