{"report":"{\"v\":1,\"package\":\"atlascloud-mcp\",\"version\":\"1.7.1\",\"ranAt\":\"2026-11-16T11:46:05.791Z\",\"install\":{\"ok\":true,\"integrity\":\"sha512-z6EDvUnCu5TXZh8lRzxKXnXxaJDOy/FsE7W2d+/YVZiMBAbzI+pmBFYV05M1orS//ITRlYz0Tjk5J+kmo8RkXg==\",\"deps\":96,\"scripts\":[],\"bins\":[\"atlascloud-mcp\"],\"license\":\"MIT\",\"engines\":\"node >=18\",\"bin\":\"atlascloud-mcp\"},\"start\":{\"ok\":true,\"ms\":228,\"serverName\":\"atlascloud-mcp\",\"serverVersion\":\"1.7.1\",\"protocol\":\"2025-06-18\",\"stderr\":\"Atlas Cloud MCP Server running via stdio\\n\"},\"tools\":[{\"name\":\"atlas_search_docs\",\"description\":\"Search Atlas Cloud documentation, models, and API references by keyword.\\n\\nReturns matching models with descriptions, pricing, and links. For detailed API docs of a specific model, use atlas_get_model_info instead.\\n\\nArgs:\\n  - query (string): Search keyword to match against model names, types, providers, tags, etc.\\n\\nReturns:\\n  Markdown-formatted list of matching models with key information.\\n\\nExamples:\\n  - \\\"video generation\\\" -> finds all video generation models\\n  - \\\"deepseek\\\" -> finds all DeepSeek models\\n  - \\\"image edit\\\" -> finds image editing models\\n  - \\\"qwen\\\" -> finds all Qwen models\",\"params\":[\"query\"]},{\"name\":\"atlas_list_models\",\"description\":\"List available models on Atlas Cloud, optionally filtered.\\n\\nThere are 400+ models, so a listing is capped: the response says how many matched and how many are shown. Narrow it with type / kind / query rather than raising limit, otherwise the tail is cut off.\\n\\nArgs:\\n  - type (string, optional): \\\"Text\\\", \\\"Image\\\", \\\"Video\\\" or \\\"Audio\\\"\\n  - kind (string, optional): Sub-kind that cuts across type: \\\"3d\\\", \\\"tts\\\", \\\"stt\\\", \\\"music\\\", \\\"lyrics\\\"\\n  - query (string, optional): Keyword matched against model ID, name, provider and tags\\n  - limit (number, optional): Max rows, 1-200. Default 60\\n\\nReturns:\\n  Markdown list grouped by type, one line per model: model ID, display name, kind, provider.\\n\\nType notes:\\n  - image-to-3D and text-to-3D models are Image-type (use kind=\\\"3d\\\" to isolate them)\\n  - Audio-type covers text-to-speech, music generation, lyrics generation and speech-to-text/ASR\\n  - lipsync / talking-avatar models are Video-type\\n\\nExamples:\\n  - type=\\\"Image\\\" -> image generation models (includes 3D)\\n  - kind=\\\"stt\\\" -> speech-to-text models only\\n  - kind=\\\"music\\\" -> music generation models (Suno, MiniMax Music)\\n  - query=\\\"kling\\\", type=\\\"Video\\\" -> Kling video models\\n  - No params -> the first 60 models across all types\",\"params\":[\"type\",\"kind\",\"query\",\"limit\"]},{\"name\":\"atlas_get_model_info\",\"description\":\"Get detailed information about a specific Atlas Cloud model: metadata, pricing in that model's own billing unit, full input/output schema, and usage examples.\\n\\nFor media models (image / video / audio / 3D) this renders the model's OpenAPI schema with every accepted parameter and a two-step cURL example. For LLM models it renders the endpoint and request body for the protocol that model actually speaks (OpenAI chat completions, OpenAI Responses, Anthropic Messages or native Gemini), plus where to read the reply from.\\n\\nArgs:\\n  - model (string): The model ID (e.g. \\\"deepseek-ai/deepseek-v3.2\\\", \\\"kwaivgi/kling-v3.0-std/text-to-video\\\")\\n\\nReturns:\\n  Markdown-formatted model details including:\\n  - Model metadata (type, kind, provider, input/output modalities, context length)\\n  - Pricing with the correct unit (per image / second / 1K characters / minute of audio / generation)\\n  - Full API input/output schema with parameter descriptions, or the chat protocol contract for LLMs\\n  - Required and optional parameters with defaults\\n  - cURL usage examples and the playground link\\n\\nExamples:\\n  - model=\\\"deepseek-ai/deepseek-v3.2\\\" -> DeepSeek V3.2 details and chat API contract\\n  - model=\\\"suno/chirp-v5\\\" -> Suno music model parameters and output shape\\n  - model=\\\"kwaivgi/kling-v3.0-std/text-to-video\\\" -> Kling video model API docs\",\"params\":[\"model\"]},{\"name\":\"atlas_generate_image\",\"description\":\"Generate an image using Atlas Cloud API. This also covers image editing and image-to-3D / text-to-3D models (3D models are Image-type and return mesh files such as GLB/OBJ/USDZ).\\n\\nThis tool submits the generation request and returns immediately with a prediction ID. Use atlas_get_prediction to check the result later.\\n\\nWhat the result looks like:\\n  - Image models: one or more image URLs (Midjourney-style models return 4 at once)\\n  - 3D models: several files in one result — the mesh (.glb), a packaged archive (.zip) and a preview image\\n\\nParameters are validated against the model's schema BEFORE the request is submitted: if a parameter is missing, has the wrong type, or is not accepted, the tool returns a precise error and does NOT spend credits.\\n\\nIMPORTANT: The \\\"model\\\" parameter requires an exact model ID (e.g., \\\"bytedance/seedream-v5.0-pro/text-to-image\\\"). If you don't know the exact model ID, you MUST first call atlas_list_models with type=\\\"Image\\\" to find it. Do NOT guess model IDs.\\n\\nYou should also use atlas_get_model_info to understand what parameters a specific image model accepts before calling this tool.\\n\\nArgs:\\n  - model (string, required): The exact image model ID. Use atlas_list_models to find valid IDs.\\n  - params (object, required): Model-specific parameters as a JSON object. Each model has different parameters defined in its schema. Common params include \\\"prompt\\\", \\\"image_size\\\", \\\"num_inference_steps\\\", etc. Input images go in a field whose name and shape vary by model — often \\\"image\\\" (string) or \\\"images\\\" (array of URLs), sometimes \\\"first_frame_image\\\" or \\\"reference_images\\\". Use atlas_get_model_info to see the full parameter list for your chosen model, and atlas_upload_media to turn a local file into a URL.\\n  - dry_run (boolean, optional): Build and validate the request, show the exact body that would be sent, and stop. Nothing is submitted and no credits are spent. Use this to check what a call will do before paying for it.\\n\\nReturns:\\n  A prediction ID to check the result with atlas_get_prediction.\\n\\nExamples:\\n  - model=\\\"bytedance/seedream-v5.0-pro/text-to-image\\\", params={\\\"prompt\\\": \\\"a cat in space\\\"}\\n  - model=\\\"alibaba/qwen-image/text-to-image-plus\\\", params={\\\"prompt\\\": \\\"sunset over mountains\\\", \\\"image_size\\\": \\\"1024x1024\\\"}\\n  - model=\\\"alibaba/wan-2.5/image-edit\\\", params={\\\"prompt\\\": \\\"make it snow\\\", \\\"images\\\": [\\\"https://example.com/photo.png\\\"]}\\n  - model=\\\"tencent/hunyuan3d-rapid/image-to-3d\\\", params={\\\"image\\\": \\\"https://example.com/photo.jpg\\\", \\\"format\\\": \\\"GLB\\\"}\",\"params\":[\"model\",\"params\",\"dry_run\"]},{\"name\":\"atlas_generate_video\",\"description\":\"Generate a video using Atlas Cloud API. This covers text-to-video, image-to-video, video-to-video / video editing, and lipsync / talking-avatar models.\\n\\nThis tool submits the generation request and returns immediately with a prediction ID. Use atlas_get_prediction to check the result later.\\n\\nWhat the result looks like: one video file URL (some models also report the output duration).\\n\\nParameters are validated against the model's schema BEFORE the request is submitted: if a parameter is missing, has the wrong type, or is not accepted, the tool returns a precise error and does NOT spend credits.\\n\\nIMPORTANT: The \\\"model\\\" parameter requires an exact model ID (e.g., \\\"kwaivgi/kling-v3.0-std/text-to-video\\\"). If you don't know the exact model ID, you MUST first call atlas_list_models with type=\\\"Video\\\" to find it. Do NOT guess model IDs.\\n\\nYou should also use atlas_get_model_info to see the full parameter list and schema for your chosen video model before calling this tool.\\n\\nArgs:\\n  - model (string, required): The exact video model ID. Use atlas_list_models to find valid IDs.\\n  - params (object, required): Model-specific parameters as a JSON object. Parameters vary by model - use atlas_get_model_info to see available params. Common ones include:\\n    - \\\"prompt\\\" (string): Text description of the video. A few multi-shot models take an array of prompts instead.\\n    - \\\"image\\\" / \\\"image_url\\\" / \\\"images\\\" / \\\"first_frame_image\\\": Source image(s) for image-to-video models — the field name and whether it is a string or an array vary by model\\n    - \\\"video\\\" / \\\"video_url\\\" / \\\"video_clips\\\": Source footage for video-to-video, editing and extension models\\n    - \\\"audio_url\\\" (string): Speech track for lipsync / talking-avatar models\\n    - \\\"duration\\\" (number): Video duration in seconds\\n    - \\\"aspect_ratio\\\" / \\\"ratio\\\" (string): e.g., \\\"16:9\\\", \\\"9:16\\\"\\n  Use atlas_upload_media to turn a local file into a URL first.\\n  - dry_run (boolean, optional): Build and validate the request, show the exact body that would be sent, and stop. Nothing is submitted and no credits are spent. Use this to check what a call will do before paying for it.\\n\\nReturns:\\n  A prediction ID to check the result with atlas_get_prediction. Video generation typically takes 1-5 minutes.\\n\\nExamples:\\n  - model=\\\"kwaivgi/kling-v3.0-std/text-to-video\\\", params={\\\"prompt\\\": \\\"a rocket launching into space\\\", \\\"duration\\\": 5}\\n  - model=\\\"bytedance/seedance-2.5/image-to-video\\\", params={\\\"prompt\\\": \\\"camera panning right\\\", \\\"image_url\\\": \\\"https://example.com/photo.jpg\\\"}\\n  - model=\\\"alibaba/wan-2.6/video-to-video\\\", params={\\\"prompt\\\": \\\"restyle as winter\\\", \\\"video\\\": \\\"https://example.com/clip.mp4\\\"}\",\"params\":[\"model\",\"params\",\"dry_run\"]},{\"name\":\"atlas_generate_audio\",\"description\":\"Generate audio — text-to-speech (TTS), music, and song lyrics — using Atlas Cloud API.\\n\\nThis covers the generative Audio-type models:\\n  - TTS / voice models (e.g. \\\"bytedance/seed-audio-1.0\\\", \\\"minimax/speech-2.6-hd\\\", \\\"elevenlabs/v3/text-to-speech\\\", \\\"google/gemini-2.5-pro-tts\\\", \\\"xai/tts-v1\\\")\\n  - Music / song models (e.g. \\\"suno/chirp-v5\\\", \\\"suno/chirp-auk\\\", \\\"suno/chirp-fenix\\\", \\\"suno/chirp-v4-5-plus\\\", \\\"minimax/music-3.0\\\") — full songs with vocals, background music, jingles\\n  - Lyrics models (e.g. \\\"minimax/lyrics-generation\\\") — these return lyrics TEXT, not audio\\n\\nThis tool submits the generation request and returns immediately with a prediction ID. Use atlas_get_prediction to check the result later.\\n\\nWhat the result looks like:\\n  - TTS: one audio file URL\\n  - Suno music models: TWO audio tracks per generation, plus a cover-art image\\n  - MiniMax music: one audio file URL\\n  - Lyrics models: the lyrics text itself, plus a song title and style tags — nothing to download\\n\\nParameters are validated against the model's schema BEFORE the request is submitted: if a parameter is missing, has the wrong type, or is not accepted, the tool returns a precise error and does NOT spend credits.\\n\\nIMPORTANT: The \\\"model\\\" parameter requires an exact model ID (e.g., \\\"bytedance/seed-audio-1.0\\\"). If you don't know the exact model ID, call atlas_list_models with type=\\\"Audio\\\" (or kind=\\\"tts\\\" / kind=\\\"music\\\") to find it. Do NOT guess model IDs.\\n\\nNOTE: For speech-to-text (transcription / ASR), use atlas_transcribe_audio instead.\\n\\nYou should also use atlas_get_model_info to see the full parameter list and schema for your chosen audio model before calling this tool. Different models accept different params (voice/speaker IDs, format, sample rate, speed, lyrics, style tags, custom vs inspiration mode, etc.).\\n\\nArgs:\\n  - model (string, required): The exact audio model ID. Use atlas_list_models with type=\\\"Audio\\\" to find valid IDs.\\n  - params (object, required): Model-specific parameters as a JSON object. For TTS the main field is usually \\\"text\\\" (the content to synthesize); for music models it is usually \\\"prompt\\\" and/or \\\"lyrics\\\"/\\\"style\\\". Use atlas_get_model_info to see available params.\\n  - dry_run (boolean, optional): Build and validate the request, show the exact body that would be sent, and stop. Nothing is submitted and no credits are spent. Use this to check what a call will do before paying for it.\\n\\nReturns:\\n  A prediction ID to check the result with atlas_get_prediction.\\n\\nExamples:\\n  - model=\\\"bytedance/seed-audio-1.0\\\", params={\\\"text\\\": \\\"Welcome to Atlas Cloud.\\\"}\\n  - model=\\\"bytedance/seed-audio-1.0\\\", params={\\\"text\\\": \\\"Hello there.\\\", \\\"format\\\": \\\"mp3\\\", \\\"sample_rate\\\": 24000}\\n  - model=\\\"suno/chirp-v5\\\", params={\\\"prompt\\\": \\\"upbeat synthwave song about coding at night\\\"}\\n  - model=\\\"suno/chirp-v5\\\", params={\\\"custom\\\": true, \\\"prompt\\\": \\\"[Verse]\\\\n...\\\", \\\"title\\\": \\\"Midnight Drive\\\", \\\"style\\\": \\\"synthwave, female vocal\\\"}\\n  - model=\\\"minimax/music-3.0\\\", params={\\\"prompt\\\": \\\"gentle acoustic guitar background musi\",\"params\":[\"model\",\"params\",\"dry_run\"]},{\"name\":\"atlas_transcribe_audio\",\"description\":\"Transcribe speech audio to text (ASR / speech-to-text) using Atlas Cloud API.\\n\\nUse this for: transcribing recordings, meetings, interviews, podcasts, voice notes; getting subtitles/captions text from audio.\\n\\nThis tool submits the transcription request and returns immediately with a prediction ID. Use atlas_get_prediction to check the result later.\\n\\nWhat the result looks like: the transcript TEXT is returned directly in the output — there is no file to download. When the model supports it, the result also carries the audio duration, per-word timestamps and speaker labels.\\n\\nParameters are validated against the model's schema BEFORE the request is submitted: if a parameter is missing, has the wrong type, or is not accepted, the tool returns a precise error and does NOT spend credits.\\n\\nIMPORTANT: The \\\"model\\\" parameter requires an exact speech-to-text model ID (e.g., \\\"bytedance/seed-asr-2.0\\\", \\\"xai/stt-v1\\\"). Find them with atlas_list_models kind=\\\"stt\\\". Do NOT guess model IDs — passing a text-to-speech model here is rejected before it costs anything.\\n\\nThe audio must be reachable via URL. For local files, first call atlas_upload_media to get a URL.\\n\\nArgs:\\n  - model (string, required): The exact speech-to-text model ID.\\n  - params (object, required): Model-specific parameters. The main field is usually \\\"audio_url\\\" (URL of the audio to transcribe). Other common params: \\\"language\\\", \\\"format\\\", \\\"enable_punc\\\", \\\"show_utterances\\\". Use atlas_get_model_info to see available params.\\n  - dry_run (boolean, optional): Build and validate the request, show the exact body that would be sent, and stop. Nothing is submitted and no credits are spent. Use this to check what a call will do before paying for it.\\n\\nReturns:\\n  A prediction ID to check the result with atlas_get_prediction. The output is the transcribed text.\\n\\nExamples:\\n  - model=\\\"bytedance/seed-asr-2.0\\\", params={\\\"audio_url\\\": \\\"https://example.com/meeting.mp3\\\"}\\n  - model=\\\"bytedance/seed-asr-2.0\\\", params={\\\"audio_url\\\": \\\"https://example.com/interview.wav\\\", \\\"show_utterances\\\": true}\",\"params\":[\"model\",\"params\",\"dry_run\"]},{\"name\":\"atlas_chat\",\"description\":\"Send a chat request to an LLM on Atlas Cloud.\\n\\nThe endpoint and request format are chosen automatically from the model's declared protocol — OpenAI chat completions, OpenAI Responses, Anthropic Messages, or native Gemini generateContent. You do not need to know which one a model uses.\\n\\nMultimodal input is supported for models that accept it: attach image/video/audio URLs (or data: URIs) to a message and they are converted to the right shape for that model's protocol. Use atlas_get_model_info to see a model's input modalities.\\n\\nArgs:\\n  - model (string, required): The LLM model ID (e.g. \\\"deepseek-ai/deepseek-v3.2\\\", \\\"google/gemini-3.1-flash-lite\\\")\\n  - messages (array, required): Message objects with \\\"role\\\" (\\\"system\\\" | \\\"user\\\" | \\\"assistant\\\") and \\\"content\\\" (text).\\n    Optional per message: \\\"images\\\", \\\"videos\\\", \\\"audios\\\" — arrays of URLs or data: URIs.\\n  - temperature (number, optional): Sampling temperature, 0-2.\\n  - max_tokens (number, optional): Maximum tokens in the response.\\n  - top_p (number, optional): Nucleus sampling, 0-1.\\n  - extra_params (object, optional): Additional sampling parameters passed through unchanged\\n    (e.g. {\\\"top_k\\\": 40, \\\"seed\\\": 7, \\\"stop\\\": [\\\"\\\\n\\\\n\\\"], \\\"frequency_penalty\\\": 0.2}).\\n    Check atlas_get_model_info for the parameters a model supports. Ignored on the Gemini protocol.\\n\\nReturns:\\n  The generated message, finish reason and token usage.\\n\\nExamples:\\n  - model=\\\"deepseek-ai/deepseek-v3.2\\\", messages=[{\\\"role\\\": \\\"user\\\", \\\"content\\\": \\\"Hello\\\"}]\\n  - model=\\\"qwen/qwen3.7-plus\\\", messages=[{\\\"role\\\": \\\"system\\\", \\\"content\\\": \\\"You are a helpful assistant\\\"}, {\\\"role\\\": \\\"user\\\", \\\"content\\\": \\\"Explain quantum computing\\\"}], temperature=0.7\\n  - model=\\\"google/gemini-3.1-flash-lite\\\", messages=[{\\\"role\\\": \\\"user\\\", \\\"content\\\": \\\"What is in this picture?\\\", \\\"images\\\": [\\\"https://example.com/photo.jpg\\\"]}]\",\"params\":[\"model\",\"messages\",\"temperature\",\"max_tokens\",\"top_p\",\"extra_params\"]},{\"name\":\"atlas_get_prediction\",\"description\":\"Check the status and result of an Atlas Cloud generation task (image, video, 3D, audio/TTS, music, or speech-to-text).\\n\\nUse this after submitting a generation request to check if the result is ready.\\n\\nIf the status is still \\\"processing\\\", \\\"created\\\", \\\"starting\\\" or \\\"queued\\\", wait a moment and try again. \\\"completed\\\"/\\\"succeeded\\\" means success; \\\"failed\\\", \\\"timeout\\\" and \\\"canceled\\\" are final — do not keep polling those.\\n\\nIMPORTANT: the result is not always a file URL. Depending on the model it can be:\\n  - Media URLs (image / video / audio / 3D files) — show them to the user and offer to download them\\n  - Plain TEXT returned directly in the output (speech-to-text transcripts, generated lyrics) — this IS the content; do not try to download it\\n  - Extra structured data: transcript timing info, lyrics title/style tags, or the cover art that ships with a generated song\\n\\nArgs:\\n  - prediction_id (string, required): The prediction ID returned from a generation request\\n\\nReturns:\\n  Status, cost, and the result rendered according to what the model actually returned.\\n\\nExamples:\\n  - prediction_id=\\\"6fca9369e2fa4998a0fcad1e61bdd30a\\\" -> check generation status\",\"params\":[\"prediction_id\"]},{\"name\":\"atlas_list_predictions\",\"description\":\"List past generation tasks (image, video, 3D, audio, speech-to-text) made with this API key, newest first.\\n\\nUse this to:\\n  - Recover a prediction ID that was lost (e.g. the submit call timed out but the task was created and is still billed)\\n  - Review what was generated earlier and fetch the output URLs again\\n  - Check which recent tasks failed and why\\n\\nEach row shows the prediction ID, time, model, status and a short preview of the result. Pass the ID to atlas_get_prediction for the full result.\\n\\nArgs:\\n  - prediction_id (string, optional): Look up one exact task by ID.\\n  - model (string, optional): Filter by exact model ID, e.g. \\\"bytedance/seed-asr-2.0\\\".\\n  - status (string, optional): Filter by status: \\\"created\\\", \\\"processing\\\", \\\"completed\\\", \\\"failed\\\", \\\"timeout\\\".\\n  - limit (number, optional): Rows per page, 1-50. Default 10.\\n  - page (number, optional): 1-based page number. Default 1.\\n\\nReturns:\\n  A Markdown list of past tasks with IDs, statuses and result previews.\\n\\nExamples:\\n  - (no params) -> the 10 most recent tasks\\n  - status=\\\"failed\\\" -> recent failures with their error messages\\n  - model=\\\"suno/chirp-v5\\\", limit=5 -> the last 5 Suno songs\",\"params\":[\"prediction_id\",\"model\",\"status\",\"limit\",\"page\"]},{\"name\":\"atlas_quick_generate\",\"description\":\"One-step image, video, or audio generation - automatically finds the model by keyword, fetches its schema, builds parameters, and submits the task.\\n\\nCovers all generation tasks: text-to-image, image editing, 3D (image/text-to-3D), text/image/video-to-video, lipsync & talking-avatar, TTS, music, lyrics and speech-to-text.\\n\\nMedia inputs are mapped onto whatever field the chosen model actually declares (image / images / image_url / first_frame_image / video_url / audio_url / ...), including array-typed fields. If a model has no field for what you passed, the response says so instead of failing silently.\\n\\nParameters are validated against the model's schema BEFORE submitting. If extra_params contains fields the model does not accept (or wrong values), the tool returns a precise error and does NOT spend credits.\\n\\nIMPORTANT: If this tool fails to find a model, call atlas_list_models first to get the exact model list, then use atlas_generate_image / atlas_generate_video / atlas_generate_audio with the exact model ID instead. Do NOT invent extra_params - only pass parameters you know the model accepts (check atlas_get_model_info).\\n\\nArgs:\\n  - model_keyword (string, required): A keyword to search for the model. Use the model's display name or key words (e.g., \\\"Nano Banana\\\", \\\"Seedream\\\", \\\"Kling\\\", \\\"Vidu\\\", \\\"Seedance\\\", \\\"Seed Audio\\\", \\\"Suno\\\", \\\"Omni Human\\\")\\n  - type (string, required): Generation type: \\\"Image\\\", \\\"Video\\\", or \\\"Audio\\\"\\n  - prompt (string, required): Text description of what to generate (for TTS, the text to synthesize; for music, the song description). Some models take no prompt at all (upscalers, lipsync, transcription) — pass a short description anyway and the response will note that it was ignored.\\n  - image_url (string or string[], optional): Source image(s) for image editing, image-to-video, image-to-3D, reference images or talking-avatar models. Pass an array for models that take multiple images — order matters: for start-end-frame video models the first URL becomes the opening frame and the second the closing frame.\\n  - video_url (string, optional): Source video for video-to-video, video editing or video-extension models\\n  - audio_url (string, optional): Source audio for lipsync / talking-avatar video models (the speech the character should say) or for speech-to-text models\\n  - extra_params (object, optional): Additional model-specific parameters to override defaults (e.g., {\\\"duration\\\": 10, \\\"aspect_ratio\\\": \\\"16:9\\\"}). Only include parameters the model's schema accepts.\\n  - dry_run (boolean, optional): Resolve the model, build the request and validate it, then show the exact body that would be sent and stop. Nothing is submitted and no credits are spent. Useful when you are unsure which field a media URL will land on, or which model a keyword resolves to.\\n\\nReturns:\\n  A prediction ID to check the result with atlas_get_prediction.\\n\\nExamples:\\n  - model_keyword=\\\"nano banana\\\", type=\\\"Image\\\", prompt=\\\"a cute cat in space\\\"\\n  - model_keyword=\\\"seedream v5\\\", ty\",\"params\":[\"model_keyword\",\"type\",\"prompt\",\"image_url\",\"video_url\",\"audio_url\",\"extra_params\",\"dry_run\"]},{\"name\":\"atlas_upload_media\",\"description\":\"Upload a local file to Atlas Cloud and get a publicly accessible URL.\\n\\nUse this whenever a model needs an input file but you only have a local path. Models take inputs by URL, never as raw bytes.\\n\\nSupported file types:\\n  - Images (jpg, png, webp, gif, bmp, avif) — image editing, image-to-video, image-to-3D, reference images, vision chat\\n  - Audio (mp3, wav, m4a, aac, ogg, flac) — speech-to-text, lipsync / talking-avatar, voice cloning references\\n  - Video (mp4, mov, webm) — video-to-video, video editing, video understanding\\n  - Documents (pdf, docx, xlsx, pptx, txt, md) — models whose schema declares a document input field\\n\\nWorkflow:\\n  1. Upload the local file with this tool to get a URL\\n  2. Pass that URL to the parameter the model expects. The field name varies by model — check atlas_get_model_info. Common ones:\\n     - images (array) / image / image_url / first_frame_image / reference_images  for pictures\\n     - audio_url / audio / reference_audio                                        for sound\\n     - video / video_url / reference_videos                                       for footage\\n     - file                                                                       for documents\\n\\nNote: the file extension decides how the file is classified downstream, so keep the real extension on the local path — a .png that actually contains AVIF data will be rejected by the model, not here.\\n\\nIMPORTANT: This upload is intended for temporary use with Atlas Cloud generation tasks only. Uploaded files may be cleaned up periodically. Do NOT use this as a permanent file hosting service. Abuse (e.g. bulk uploads unrelated to generation tasks) may result in API key suspension.\\n\\nArgs:\\n  - file_path (string, required): Absolute path to the local file to upload\\n\\nReturns:\\n  The publicly accessible download URL of the uploaded file.\\n\\nExamples:\\n  - file_path=\\\"/Users/me/photos/cat.jpg\\\" -> a URL to pass as image / images[0]\\n  - file_path=\\\"/Users/me/recordings/meeting.mp3\\\" -> a URL to pass as audio_url for speech-to-text\",\"params\":[\"file_path\"]},{\"name\":\"atlas_get_balance\",\"description\":\"Get the current Atlas Cloud account balance and credit summary for the authenticated API key.\\n\\nReturns available balance, cash, bonus, frozen amount, and credit grant status.\\n\\nArgs:\\n  None.\\n\\nExamples:\\n  - (no params) -> current account balance\",\"params\":[]},{\"name\":\"atlas_get_model_usage\",\"description\":\"Get daily model usage buckets (requests, tokens, image/video counts) for the authenticated API key over a UTC date range.\\n\\nOmitting group_by returns total daily buckets. Use group_by to break usage down by model_type, model, and/or api_key.\\n\\nArgs:\\n  - start_date (string, required): Inclusive UTC start date, YYYY-MM-DD.\\n  - end_date (string, required): Exclusive UTC end date, YYYY-MM-DD (range <= 180 days).\\n  - scope (string, optional): \\\"self\\\" (default) or \\\"account\\\".\\n  - group_by (array, optional): Any of \\\"model_type\\\", \\\"model\\\", \\\"api_key\\\".\\n  - model_types (array, optional): Filter by model type, e.g. [\\\"text\\\",\\\"image\\\"].\\n  - model_ids (array, optional): Filter by model ID (<= 100).\\n  - limit (number, optional): Max grouped rows (1-1000, default 100).\\n  - page (string, optional): Pagination cursor from a previous next_page.\\n\\nExamples:\\n  - start_date=\\\"2026-06-01\\\", end_date=\\\"2026-06-08\\\"\\n  - start_date=\\\"2026-06-01\\\", end_date=\\\"2026-07-01\\\", group_by=[\\\"model\\\"]\",\"params\":[\"start_date\",\"end_date\",\"scope\",\"group_by\",\"model_types\",\"model_ids\",\"limit\",\"page\"]},{\"name\":\"atlas_get_model_costs\",\"description\":\"Get daily model cost buckets (spend amount) for the authenticated API key over a UTC date range.\\n\\nOmitting group_by returns total daily buckets. Use group_by to break costs down by model_type, model, and/or api_key.\\n\\nArgs:\\n  - start_date (string, required): Inclusive UTC start date, YYYY-MM-DD.\\n  - end_date (string, required): Exclusive UTC end date, YYYY-MM-DD (range <= 180 days).\\n  - scope (string, optional): \\\"self\\\" (default) or \\\"account\\\".\\n  - group_by (array, optional): Any of \\\"model_type\\\", \\\"model\\\", \\\"api_key\\\".\\n  - model_types (array, optional): Filter by model type, e.g. [\\\"text\\\",\\\"image\\\"].\\n  - model_ids (array, optional): Filter by model ID (<= 100).\\n  - limit (number, optional): Max grouped rows (1-1000, default 100).\\n  - page (string, optional): Pagination cursor from a previous next_page.\\n\\nExamples:\\n  - start_date=\\\"2026-06-01\\\", end_date=\\\"2026-07-01\\\"\\n  - start_date=\\\"2026-06-01\\\", end_date=\\\"2026-07-01\\\", group_by=[\\\"model\\\"]\",\"params\":[\"start_date\",\"end_date\",\"scope\",\"group_by\",\"model_types\",\"model_ids\",\"limit\",\"page\"]}],\"toolsHash\":\"b515941bb7c342860b732aa9e996a3957ed6f0cf31d6230a2509f9ff2decbaff\",\"wrote\":[],\"callsAt\":\"2026-11-16T11:46:12.110Z\",\"calls\":[{\"tool\":\"atlas_list_models\",\"ok\":true,\"ms\":570,\"bytes\":4964},{\"tool\":\"atlas_list_predictions\",\"ok\":false,\"ms\":2,\"bytes\":0,\"error\":\"tool reported an error: Error: ATLASCLOUD_API_KEY is not set. Please add it to your MCP configuration:\\n\\n{\\n  \\\"mcpServers\\\": {\\n    \\\"atlascloud\\\": {\\n      \\\"command\\\": \\\"npx\\\",\\n      \\\"args\\\": [\\\"-y\\\", \\\"atlascloud-mcp\\\"],\\n      \\\"env\\\": {\\n \"},{\"tool\":\"atlas_get_balance\",\"ok\":false,\"ms\":1,\"bytes\":0,\"error\":\"tool reported an error: Error: ATLASCLOUD_API_KEY is not set. Please add it to your MCP configuration:\\n\\n{\\n  \\\"mcpServers\\\": {\\n    \\\"atlascloud\\\": {\\n      \\\"command\\\": \\\"npx\\\",\\n      \\\"args\\\": [\\\"-y\\\", \\\"atlascloud-mcp\\\"],\\n      \\\"env\\\": {\\n \"}],\"host\":{\"commit\":\"eaaffffaf52255d1b77dc4aacffdfaf353f0c12c\",\"run\":\"https://github.com/itsHaddad/bragproof/actions/runs/43857070216\"},\"network\":{\"names\":[\"registry.npmjs.org\"],\"conns\":[],\"callNames\":[\"api.atlascloud.ai\"],\"callConns\":[]}}","token":"eyJhbGciOiJSUzI1NiIsImtpZCI6IjM4ODI2YjE3LTZhMzAtNWY5Yi1iMTY5LThiZWI4MjAyZjcyMyIsInR5cCI6IkpXVCIsIng1dCI6InlrTmFZNHFNX3RhNGsyVGdaT0NFWUxrY1lsQSJ9.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.nPXs79_-Yzn1K4LRKpAWqNUFe4DTLAGpEV3Qslh2lY_4EWA7FiWJ5ZNfFA7JNwzPgbmvxnMrcte0NNnXoUH9kLohnEE66Wh0ocpY8R_A88WQ-_QMAKpTHUa63CvOs9uIPuNelDKTgp8fboycLMjSa7DCrXeTxBN5ZjRb3oZhdvxj5Mj9vNz11cbmlzjlYTDYxBYt4fToMIvwZtKonAYYqeefIUUAbukM9fJqDnwjZYchBabE3uFSPey8RYe-7VIV4GB6CUpAe3x_MBINV9y5ijoIl5ZSEK3BFi3elew1fBqgUAqeJcsGXgbzmyp4Zd9eOVqumYBv1nW4BCmhk6VfEg"}