> ## Documentation Index > Fetch the complete documentation index at: https://docs.tensorpix.ai/llms.txt > Use this file to discover all available pages before exploring further. # Enhance uploaded video > Enhance the uploaded video. To upload the video, use the [upload videos endpoint](https://backend.tensorpix.ai/api/schema/redoc/#tag/videos/operation/videos_create). After the video is uploaded, use this endpoint to start the video enhancing job. Make sure to set the `input_video` attribute to the uploaded video's ID which is returned in the response of the [upload videos endpoint](https://backend.tensorpix.ai/api/schema/redoc/#tag/videos/operation/videos_create). ## AI filters AI filters are called `ml-models` in the API docs as they are actually machine learning models. See the `/api/ml-models` to learn more about ML models selection. After choosing the appropriate models, pass the list of their IDs in the `ml_models` parameter. You should pass at least one AI filter. ## Free Preview Jobs Free Previews will process a short segment of the video (10sec) and won't cost any 🪙 Credits. To enable Free Previews in a job, set the `preview = True`. You can use the `start_frame` to set the starting frame of the Free Preview segment. `end_frame` parameter is ignored if `preview = True`. ## Enhancing specific video segments To process a whole video, don't specify `start_frame` and `end_frame` and set `preview = False`. To process a specific video segment, set the `start_frame` to the starting frame of the segment, and `end_frame` to the end frame of the segment. `end_frame` must be > `start_frame`. ## Stabilization `stabilization_smoothing` parameter is ignored unless the AI Stabilization Filter is set as one of the ML models. ## Sharpening `sharpen_strength` parameter is ignored unless the AI Sharpen Filter is set as one of the ML models. ## Post enhancing clean-up > **⚠️ IMPORTANT**: Your account will be limited by the total storage of all enhanced and uploaded videos. While this limit will be set to a high number (1TB+), we highly recommend you delete the enhanced video after the enhanced video is copied to your server. This ensures you won't run into storage limit errors in the future. ## OpenAPI ````yaml https://backend.tensorpix.ai/api/schema/ post /api/jobs/ openapi: 3.1.0 info: title: TensorPix OpenAPI Documentation version: 1.1.0 termsOfService: https://tensorpix.ai/terms-conditions contact: name: TensorPix Dev Team email: devs@tensorpix.ai servers: - url: https://backend.tensorpix.ai security: [] externalDocs: url: https://docs.tensorpix.ai description: Check out the full TensorPix documentation. paths: /api/jobs/: post: tags: - jobs summary: Enhance uploaded video description: >- Enhance the uploaded video. To upload the video, use the [upload videos endpoint](https://backend.tensorpix.ai/api/schema/redoc/#tag/videos/operation/videos_create). After the video is uploaded, use this endpoint to start the video enhancing job. Make sure to set the `input_video` attribute to the uploaded video's ID which is returned in the response of the [upload videos endpoint](https://backend.tensorpix.ai/api/schema/redoc/#tag/videos/operation/videos_create). ## AI filters AI filters are called `ml-models` in the API docs as they are actually machine learning models. See the `/api/ml-models` to learn more about ML models selection. After choosing the appropriate models, pass the list of their IDs in the `ml_models` parameter. You should pass at least one AI filter. ## Free Preview Jobs Free Previews will process a short segment of the video (10sec) and won't cost any 🪙 Credits. To enable Free Previews in a job, set the `preview = True`. You can use the `start_frame` to set the starting frame of the Free Preview segment. `end_frame` parameter is ignored if `preview = True`. ## Enhancing specific video segments To process a whole video, don't specify `start_frame` and `end_frame` and set `preview = False`. To process a specific video segment, set the `start_frame` to the starting frame of the segment, and `end_frame` to the end frame of the segment. `end_frame` must be > `start_frame`. ## Stabilization `stabilization_smoothing` parameter is ignored unless the AI Stabilization Filter is set as one of the ML models. ## Sharpening `sharpen_strength` parameter is ignored unless the AI Sharpen Filter is set as one of the ML models. ## Post enhancing clean-up > **⚠️ IMPORTANT**: Your account will be limited by the total storage of all enhanced and uploaded videos. While this limit will be set to a high number (1TB+), we highly recommend you delete the enhanced video after the enhanced video is copied to your server. This ensures you won't run into storage limit errors in the future. operationId: jobs_create requestBody: content: application/json: schema: $ref: '#/components/schemas/CreateJob' application/x-www-form-urlencoded: schema: $ref: '#/components/schemas/CreateJob' multipart/form-data: schema: $ref: '#/components/schemas/CreateJob' required: true responses: '201': content: application/json: schema: $ref: '#/components/schemas/CreateJob' description: '' '400': content: application/json: schema: $ref: '#/components/schemas/ValidationError' description: '' '401': content: application/json: schema: $ref: '#/components/schemas/UnauthenticatedError' description: '' '403': content: application/json: schema: $ref: '#/components/schemas/ForbiddenError' description: '' '429': content: application/json: schema: $ref: '#/components/schemas/RateLimitError' description: '' '500': content: application/json: schema: $ref: '#/components/schemas/GenericError' description: '' security: - tokenAuth: [] components: schemas: CreateJob: type: object properties: id: type: integer readOnly: true codec: allOf: - $ref: '#/components/schemas/CodecEnum' description: |- Codec used for the Enhanced video. * `libx264` - H264 * `libx265` - HEVC/H265 * `prores_ks` - Apple ProRes crf: type: integer maximum: 51 minimum: 0 description: >- Constant Rate Factor (CRF) is encoding quality of the output video. 23 is a good starting value. Set to 20 or lower if you need very high quality encoding (increased file size), or higher than 25 if quality is not paramount (smaller file size). [Learn more about CRF](https://trac.ffmpeg.org/wiki/Encode/H.264#crf). qscale: type: integer maximum: 32 minimum: 0 description: >- Represents the `qscale` parameter in ffmpeg. Used with ProRes codec. Set the value between 9 and 13 for optimal results. [Learn more about Qscale](https://trac.ffmpeg.org/wiki/Encode/VFX#Prores) container: allOf: - $ref: '#/components/schemas/ContainerEnum' description: |- Container format of the Enhanced video. * `mov` - Mov * `mkv` - Mkv * `mp4` - Mp4 chroma_subsampling: oneOf: - $ref: '#/components/schemas/ChromaSubsamplingEnum' - $ref: '#/components/schemas/BlankEnum' preview: type: boolean description: >- Flag which denotes if job is free preview job or not. Free preview jobs are not charged but the output video is watermarked and only a short video segment is enhanced. start_frame: type: - integer - 'null' maximum: 2147483647 minimum: 0 description: Cuts the video starting from this frame before enhancing it. end_frame: type: - integer - 'null' maximum: 2147483647 minimum: 1 description: Cuts the video ending on this frame before enhancing it. input_video: type: - integer - 'null' output_video: type: - integer - 'null' description: >- Once the job is finished, the `output_video` attribute will be populated with the enhanced video metadata. If the job is canceled or it fails, the output_video attribute will remain null. owner: type: string format: email title: Email address readOnly: true ml_models: type: array items: type: integer output_resolution: allOf: - $ref: '#/components/schemas/OutputResolutionEnum' minimum: -2147483648 maximum: 2147483647 grain: type: number format: double maximum: 1 minimum: 0 description: >- Grain amount to add to the enhanced video in range [0, 1]. 0 is default. comparison: type: boolean description: >- Flag which denotes if the output video will be side by side video of the original and enhanced video. stabilization_smoothing: type: integer maximum: 40 minimum: 0 description: >- If AI Stabilization filter is enabled, this parameter controls the intensity of stabilization. Ignored if AI Stabilization is disabled. Use higher values for very shaky videos, and lower strength for less shaky videos. Video may be zoomed in if the video is very shaky. target_fps: type: number format: double maximum: 240 minimum: 1 description: >- Target FPS for the enhanced video. If not provided, the original FPS will be used. If provided, should be greater than or equal to the original FPS. If lower, it will be set to the original FPS. **IMPORTANT**: Won't be applied if the FPS BOOST model isn't passed in `ml_models.` playback_speed: type: number format: double maximum: 1 minimum: 0.1 description: >- Playback speed of the video. 1.0 means the video will play at normal speed. 0.5 means the video will play at half speed with AI interpolated frames in between. **IMPORTANT**: Won't be applied if the SLOW MOTION model isn's passed in `ml_models.` sharpen_strength: type: number format: double maximum: 10 minimum: 0 description: >- If AI Sharpen filter is enabled, this parameter controls the intensity of sharpening. Ignored if AI Sharpen filter is disabled. cost_usd: type: string format: decimal pattern: ^-?\d{0,6}(?:\.\d{0,4})?$ readOnly: true description: Job cost in 🪙 Credits. prores_profile: allOf: - $ref: '#/components/schemas/ProresProfileEnum' minimum: -2147483648 maximum: 2147483647 required: - cost_usd - id - ml_models - owner ValidationError: type: object properties: errors: type: object additionalProperties: type: array items: type: string non_field_errors: type: array items: type: string required: - errors - non_field_errors UnauthenticatedError: type: object properties: detail: type: string required: - detail ForbiddenError: type: object properties: detail: type: string required: - detail RateLimitError: type: object properties: detail: type: string required: - detail GenericError: type: object properties: detail: type: string required: - detail CodecEnum: enum: - libx264 - libx265 - prores_ks type: string description: |- * `libx264` - H264 * `libx265` - HEVC/H265 * `prores_ks` - Apple ProRes ContainerEnum: enum: - mov - mkv - mp4 type: string description: |- * `mov` - Mov * `mkv` - Mkv * `mp4` - Mp4 ChromaSubsamplingEnum: enum: - yuv420p - yuv422p - yuv444p - yuv420p10le - yuv422p10le - yuv444p10le type: string description: |- * `yuv420p` - 8-bit 4:2:0 * `yuv422p` - 8-bit 4:2:2 * `yuv444p` - 8-bit 4:4:4 * `yuv420p10le` - 10-bit 4:2:0 * `yuv422p10le` - 10-bit 4:2:2 * `yuv444p10le` - 10-bit 4:4:4 BlankEnum: enum: - '' OutputResolutionEnum: enum: - -1 - 640 - 720 - 854 - 1280 - 1920 - 2048 - 2560 - 3840 type: integer description: |- * `-1` - No Resize * `640` - 360p * `720` - SD * `854` - HQ * `1280` - 720p * `1920` - 1080p * `2048` - 2K * `2560` - QHD * `3840` - 4K ProresProfileEnum: enum: - -1 - 0 - 1 - 2 - 3 - 4 - 5 type: integer description: |- * `-1` - Auto * `0` - Proxy * `1` - Lt * `2` - Standard * `3` - Hq * `4` - V4444 * `5` - V4444Xq securitySchemes: tokenAuth: type: apiKey in: header name: Authorization description: Token-based authentication with required prefix "Token" ````